Structural and Functional Adaptability of Sucrose and Lactose Permeases from Escherichia coli to the Membrane Lipid Composition

Abstract

The lipid environment in which membrane proteins are embedded can influence their structure and function. Lipid−protein interactions and lipid-induced conformational changes necessary for protein function remain intractable in vivo using high-resolution techniques. Using Escherichia coli strains in which the normal phospholipid composition can be altered or foreign lipids can be introduced, we established the importance of membrane lipid composition for the proper folding, assembly, and function of E. coli lactose (LacY) and sucrose (CscB) permeases. However, the molecular mechanism underlying the lipid dependence for active transport remains unknown. Herein, we demonstrate that the structure and function of CscB and LacY can be modulated by the composition of the lipid environment. Using a combination of assays (transport activity of the substrate, protein topology, folding, and assembly into the membrane), we found that alterations in the membrane lipid composition lead to lipid-dependent structural changes in CscB and LacY. These changes affect the orientation of residues involved in LacY proton translocation and impact the rates of protonation and deprotonation of E325 by affecting the arrangement of transmembrane domains in the vicinity of the R302-E325 charge pair. Furthermore, the structural changes caused by changes in membrane lipid composition can be altered by a single-point mutation, highlighting the adaptability of these transporters to their environment. Altogether, our results demonstrate that direct interactions between a protein and its lipid environment uniquely contribute to membrane protein organization and function. Because members of the major facilitator superfamily present with well-conserved functional architecture, we anticipate that our findings can be extrapolated to other membrane protein transporters.

Graphical Abstract

Lactose permease (LacY) is a membrane protein found within the inner membrane of Escherichia coli that belongs to the major facilitator superfamily (MFS). LacY cotransports galactopyranosides and a proton into the bacterial cytoplasm and is considered a paradigm for secondary transporters in membranes. LacY is comprised of 12 transmembrane domains (TMDs) organized in two pseudosymmetrical six-α-helix bundles that surround a large hydrophilic cavity (Figure 1A). The sugar-binding site and the residues involved in proton translocation are located near the apex of the cavity in the inward-facing conformation. Extensive biophysical and biochemical studies, complemented by elucidation of several high-resolution structures, have led to an in-depth understanding of the basic mechanism of lactose/proton transport,^1–3 including the identification of six irreplaceable residues with respect to active transport and several others that are necessary for optimal sugar and proton binding and translocation.^4,5

Figure 1.

Structural and sequence conservation between LacY and CscB. The Q359H and Q353H mutations are highlighted within a black box on (A) the X-ray structure of LacY (PDB entry 2Y5Y, molecule A) and (B) the backbone of the threaded model of CscB, respectively. The structures are presented using Pymol 0.97 (DeLano Scientific, LLC). (C) Distribution of conserved charged amino acids involved in the proton wire in sugar permeases from cluster 5 of the MFS (LACY_ECOLI, lactose permease from E. coli; RAFB_ECOLX, raffinose permease from E. coli; CSCB_ECOLX, sucrose permease from E. coli; MELY_ENTCC, MelY, melibiose permease from Enterobacter cloacae subsp. cloacae; LACY_KLEOX, lactose permease from Klebsiella oxytoca; LACY_CITFR, lactose permease from Citrobacter freundii). Conserved charged residues are indicated in bold. The position of the Q → H bypass mutation is colored magenta.

Proton translocation primarily involves residues in the C-terminal six-TMD α-helical bundle, which are positioned across the cavity from the sugar-binding site. Residues Y236 (VII), D240 (VII), R302 (IX), K319 (X), H322 (X), and E325 (X) form a tightly interconnected network among TMDs VII, IX, and X in the middle of the molecule. Any disturbances in the central core of the proton translocation site may decrease sugar affinity by affecting proton binding.⁶

There is a strong correlation between the proper folding of the extramembrane domain (EMD) P7 and uphill transport function.⁷ Native folding of this domain can be assessed by recognition of EMD P7 by the conformationally sensitive monoclonal antibody 4B1 (mAb 4B1).⁷ Binding of mAb 4B1 to EMD P7 alters the pK_a of E325 in TMD X that is postulated to be involved in delivering a proton to the inside of cells coupled with substrate import.⁸ This proton dissociation appears to be the rate-limiting step in facilitated transport, but imposition of an electrochemical potential (Δμ_H⁺) across the membrane changes the rate-limiting step for active transport. This allows the permease to adopt an outward-facing conformation before the substrate, which binds to only the protonated carrier,^9,10 can be rebound from the inside. This change in the rate-limiting step is postulated to kinetically drive substrate accumulation against a concentration gradient.¹¹ Binding effects of mAb 4B1 are interpreted as a change in the interaction among residues in TMDs VII (attached to EMD P7), IX, and X, which results in a loss of active but not facilitated transport function. Because these residues are involved in proton translocation, these structural changes may explain uncoupling of transport from Δμ_H⁺. Because substrate binding affinities for either side of the membrane have been verified to be the same and independent of an imposed Δμ_H⁺,¹¹ a change in the organization of the proton wire may be sufficient to prevent coupling and active transport.

We previously demonstrated that the lipid environment is a determinant of TMD orientation for several secondary transporters of E. coli, including LacY.^12,13 When assembled in cells lacking phosphatidylethanolamine (PE), the net neutral and major phospholipid of E. coli LacY exhibits inversion of the N-terminal (NT) six-TMD α-helical bundle with respect to the plane of the membrane bilayer and the C-terminal (CT) five-TMD bundle,¹⁴ with TMD VII becoming an EMD that is exposed to the periplasm. In the absence of PE or other net neutral lipids, LacY carries out only downhill transport even though the cell maintains a robust membrane potential.^15,16 The retention of partial function indicates a compact folded structure even in the absence of PE, which was recently confirmed.¹⁷ LacY expressed in PE-lacking cells is not recognized by mAb 4B1, which can be explained by a structural change in EMD P7 due to the inverted topology and exposure to the periplasm of TMD VII. However, even when the correct native topology of LacY is maintained, partial restoration of proper EMD P7 folding is observed in cells lacking PE but containing phosphatidylcholine (PC) or the neutral glycolipid monoglucosyl diacylglycerol (MGDG).^12,13 Therefore, the misfolding of EMD P7 in the absence of PE, which is associated with a loss of or reduced active transport function, has many structural and functional parallels with the effects of binding of mAb 4B1 to EMD P7.

Why does the lipid environment affect the ability of LacY to actively transport substrate? LacY functions through conformational dynamic switches from inward-facing to outward-facing. These changes most likely involve residue−residue interactions within the transporter, specific substrate-binding and/or lipid−protein interactions. Structural transitions between inward-facing and outward-facing states may require specific lipid−protein interactions, which can be revealed only upon manipulation of the lipid composition of the membranes in which the transporter is embedded. While a complete review of the vast amount of data on kinetic analysis of LacY in whole cells, isolated right-side-out (RSO) membrane vesicles, and proteoliposomes made with various combinations of lipids is not possible, important points regarding LacY structure and function and specific lipids can be summarized as follows. Whole cells or proteoliposomes containing primarily phosphatidylglycerol (PG)/cardiolipin (CL) and lacking PE or phosphatidylserine (PS) exhibit facilitated but not active transport.¹⁵ Monomethyl-PE and dimethyl-PE, but not synthetic PC, can partially substitute for PE in supporting active transport in proteoliposomes.¹⁸ The wild type topology of LacY in the absence of PE is supported by MGDG, diglucosyl diacylglycerol (DGDG), and E. coli-made PC, both in vivo and in vitro.^19,20 However, MGDG only partially restores active transport in PE-lacking cells, while DGDG does not.²¹ Proper folding of P7 and uphill transport in cells are supported by PE and E. coli-made PC, only partially by MGDG, and not by DGDG.^20,22,23 Further dissection using LacY reconstituted in proteoliposomes demonstrated that proper folding of EMD P7 and support for uphill transport do not strictly depend on the chemical properties of the net neutral headgroup but rather on the combined physical and chemical properties of the headgroup and the fatty acids.²⁰

Other studies have attempted to identify the roles of membrane lipids in the conformational dynamics of proteins from the MFS but have so far been mostly limited to in silico and in vitro approaches. A recent study on the Salmonella typhimurium melibiose transporter MelB dissected the effects of lipid headgroups and fatty acid tails on its structure and function.²⁴ They demonstrated that PE and PG, but not CL, are required for optimal active transport in vivo using lipid-altered E. coli strains. Further dissection of substrate binding affinity, protein folding, and stability indicated that the observed lipid dependency targets the later steps of the transport process. Analysis of tightly bound lipids showed that (i) the phospholipid headgroup played little to no role in protein stability and substrate binding and (ii) fatty acid tails account for most of the direct lipid−MelB interactions, altogether highlighting the complex and multifaceted contribution of phospholipids to membrane protein structure and function. Moreover, very informative results obtained with molecular dynamics simulation (MDS) led to the identification of potential direct, PE-specific lipid−protein interactions²⁵ that could not be recapitulated with PC-containing simulated membranes. Several of these predictions were validated in vitro using purified MFS members reconstituted in nanodiscs and analyzed by the hydrogen−deuterium exchange (HDX) method,²⁵ highlighting the role specific lipid environments play in fine-tuning the structure and function of transporters. We now propose to address these questions in vivo using a set of E. coli “lipid mutants” in which the membrane lipid composition can be systematically manipulated under steady state conditions.

This study is aimed at elucidating, at the molecular level, the consequences of in vivo alterations in the membrane lipid environment on the conformational flexibility and function of LacY. The question to be addressed is whether structural changes caused by mAb 4B1 also occur with phospholipids that do not support active transport by LacY. To tackle this question, we took advantage of the fact that the sucrose permease from E. coli (CscB) exhibits the correct native topology in membranes containing or lacking PE,²⁶ while no active transport is observed without PE. CscB shares a high degree of similarity to LacY (50.5% similar and 29.5% identical) and resembles other well-studied permeases of cluster 5 of the MFS.²⁷ Homology threading of the CscB protein sequence with the structure of LacY (PDB entry 2Y5Y²⁸) as a template indicates the same type of topological organization for CscB, as well as the presence of an inward-facing hydrophilic cavity (Figure 1B). The transport activity of wild type CscB is ∼10-fold lower than that of LacY, which results in the unusual slow growth on sucrose compared to other carbon sources.²⁹ The analysis of suppressor mutants with improved growth rates led to the discovery of one specific mutation in CscB (Q353H), which results in significantly increased sucrose transport activity without a change in the expression levels of the permease.²⁹ This mutation is positioned on TMD XI, in the proximity of several of the residues involved in the putative proton wire (Figure 1C).

Because CscB and LacY have many similarities, we asked whether the increase in CscB transport activity conferred by the Q353H mutation would also be observed in LacY. We hypothesized that changes in the lipid environment restrain the conformational dynamics of LacY, which can be rescued with a single-point amino acid mutation near the proton translocation pathway. The results presented herein correlate the uncoupling of transport from Δμ_H⁺ with a lipid-dependent structural change in LacY that affects the orientation of amino acids involved in proton translocation. We now demonstrate in vivo that membrane lipids contribute to the conformational flexibility of LacY and can impact the rates of protonation and deprotonation of residue E325 by affecting the TMD arrangement in the vicinity of the R302-E325 charge pair.

MATERIALS AND METHODS

Reagents.

The detergent n-dodecyl β-d-maltoside (DDM) was purchased from Anatrace. The ECL kit, Imperial protein stain solution, horseradish peroxidase (HRP)-labeled secondary antibody, microbicinchoninic acid (BCA) protein reagent assay, 3-(N-maleimidopropionyl)biocytin (MPB), and avidin-HRP were purchased from Thermo Fisher. The QuikChange Lightning site-directed mutagenesis kit was purchased from Agilent Technologies. HiTrap columns and Vivaspin concentrators (50000 Da molecular weight cutoff) were purchased from GE Healthcare. The mouse anti-His antibody was purchased from GenScript. β-d-Galactopyranosyl-1-thio-β-d-galactopyranoside (TDG) was obtained from Cayman Chemicals. Complete protease inhibitor was purchased from Roche Applied Science. The site-directed polyclonal antibody (pAb 1043) directed against the C-terminal dodecapeptide of LacY was made by ProSci Inc. The monoclonal antibody against the LacY 4B1 epitope (mAb 4B1) and ampicillin-resistant plasmids pT7–5/lacY and pSP72/cscB expressing LacY and CscB derivatives, respectively, under OP_lac regulation were provided by H. R. Kaback (University of California, Los Angeles, Los Angeles, CA). Radiolabeled [¹⁴C]sucrose and [¹⁴C]lactose and the nonhydrolyzable lactose substrate analogue methyl-thio-β-d-galactopyranoside ([³H]TMG) were purchased from Moravek. The 1,1-methanediyl bismethanethiosulfonate (MTS-1-MTS) cross-linker was obtained from Interchim. The tobacco etch virus (TEV) protease was obtained from New England Biolabs. The protonophore carbonyl cyanide-p-trifluoromethoxyphenylhydrazone (FCCP), GF/F filters, 4-nitrophenyl-β-d-galactopyranoside (NPG), N-[4-(7-diethylamino-4-methyl-3-coumarinyl)phenyl]maleimide (CPM), and all other reagents were purchased from Sigma-Millipore.

Bacterial Strains, Plasmids, Growth Condition, and LacY Purification.

Strain AL95 (pss93::kan^R lacY::Tn9) cannot make PE and is not viable without either plasmid pDD72GM (pssA⁺ gen^R with a pSC101 temperature-sensitive replicon) or growth medium containing 50 mM MgCl₂.³⁰ Strain AL95 (grown at 37 °C) lacks PE, and strain AL95/pDD72GM (grown at 30 °C) contains the normal E. coli complement of phospholipids, including PE. Strains AD93 (pss93::kan^R) with plasmid pTMG3 (MGDG in place of PE), AD93 with plasmids pTMG3 and pACYCT7DGS (DGDG in place of PE), and AL95 with plasmid pAC-PCS_lp-Sp-Gm (PC in place of PE) were grown as previously described.^21,23 Ampicillin-resistant plasmids pT7–5/lacY and pSP72/cscB expressing LacY and CscB derivatives, respectively, under OP_lac regulation were employed. Changes in amino acid residues were accomplished by site-directed mutagenesis in derivatives of these proteins.

CscB derivatives were appended at the C-terminus with a dodecapeptide derived from the C-terminus of LacY so that the LacY-specific antibody (pAb 1043) could be used for protein detection and immunoprecipitation. LacY was engineered with a His6 tag at the C-terminus to facilitate purification. Both proteins were expressed under control of OP_lac by growth of cells in the presence of 1 mM IPTG. Cells were grown in LB rich medium containing ampicillin (100 μg/mL) as required and 50 mM MgCl₂ until an OD₆₀₀ of 0.6 was reached, induced by addition of IPTG, and grown until cell arrest occurred. Purification of LacY was carried out at 4 °C or on ice as previously described.³¹ The protein content during purification was determined by the micro-BCA protein assay, according to the manufacturer’s instructions.

Sugar Transport Assay.

For uphill transport, phospholipid-altered strains were washed once with 100 mM HEPES (pH 7.5) and 50 mM MgCl₂ and adjusted to an OD₆₀₀ of 10 (∼0.7 mg of protein/mL). Transport was initiated by the addition of 10 μL of [¹⁴C]sucrose (0.5 μCi/mL, final concentration of 0.5 mM) or [³H]TMG (0.1 μCi/mL, final concentration of 0.1 mM) to 1 mL of cells. Aliquots of 100 μL were removed at various times, quenched with 3 mL of ice-cold 100 mM HEPES (pH 7.5), 100 mM LiCl, and 50 mM MgCl₂, immediately filtered through GF/F filters, and washed using 10 mL of the same buffer. Filters were dried and counted by liquid scintillation. Uptake values were corrected for sucrose or TMG uptake by cells carrying the pSP72 vector (for sucrose transport) or pT7–5 vector (for TMG transport) only, respectively. De-energized cells were obtained by pretreating the cells with a 50 μM concentration of the protonophore FCCP for 5 min prior to assessing transport, as described above.

Counterflow Assay.

For counterflow experiments, phospholipid-altered strains were washed once with 100 mM HEPES (pH 7.5) and 50 mM MgCl₂ and adjusted to an OD₆₀₀ of 10 (∼0.7 mg of protein/mL). An aliquot (100 μL) of cells that had been concentrated and equilibrated with 0.5 mM lactose was then diluted rapidly 100-fold into 100 mM HEPES (at various pH values, ranging from 5.5 to 9.5) and 50 mM MgCl₂ containing the same concentration of [¹⁴C]lactose. The suspension was immediately mixed, and at given times, 100 μL was removed, quenched with 3 mL of ice-cold 100 mM HEPES (pH 7.5), 100 mM LiCl, and 50 mM MgCl₂, immediately filtered through GF/F filters, and washed using 10 mL of the same buffer. Filters were dried and counted by liquid scintillation. Values were corrected for lactose counterflow by cells carrying the pT7–5 vector only, as described above.

Protein Topology Mapping.

The topological orientation of LacY was determined using the substituted cysteine accessibility method as applied to transmembrane domains (SCAM), as previously described.²⁶ Using LacY with a single cysteine engineered in EMD C6 (H205C), we monitored the orientation of EMD C6 relative to the plane of the cell inner membrane.

Cross-Linking of TMDs.

RSO membrane vesicles were prepared from spheroplasts as previously described.²² Phospholipid-altered cells grown to midexponential phase were harvested, washed, and then resuspended in 10 mM potassium HEPES (pH 7.0), 0.75 M sucrose, 10 mM MgSO₄, 2.5% (w/v) LiCl, and 50 mg/mL chloramphenicol. After the addition of 1 mg/mL lysozyme, cell suspensions were chilled to 4 °C, warmed to room temperature, and subsequently incubated while being gently shaken at 30 °C for 30 min. Intact spheroplasts were collected by centrifugation (3000g for 10 min) at room temperature and resuspended at a total protein concentration of 10 mg/mL in the buffer described above without LiCl. RSO vesicles were made by diluting a pellet of spheroplasts into a 50-fold volume of 10 mM potassium HEPES buffer (pH 7.0), 10 mM MgCl₂, 1 mM DTT, and 10 mg/mL deoxyribonuclease and incubated for 10 min on ice. Intact cells and cell debris were removed by centrifugation at 4000g for 5 min, and the supernatant was centrifuged at 30000g for 30 min at 4 °C. The resulting RSO membrane vesicles were suspended to a final concentration of 2 mg of membrane proteins/mL in 50 mM potassium HEPES (pH 7.0) containing 10 mM MgCl₂. Cross-linking reactions were carried out on RSO vesicles prepared from phospholipid-altered strains at 4 °C (to minimize structural heterogeneity) for 1 h in the presence of the homobifunctional cross-linking agent MTS-1-MTS at a final concentration of 50 μM. The reactions were terminated by addition of 5 mM N-ethylmaleimide (NEM) and incubation on ice. RSO membranes were then collected by centrifugation and mixed with 20 mM Tris-HCl (pH 7.5) containing 100 mM NaCl, 2.0 mM CaCl₂, and 1% DDM. TEV protease (1 μg) was added to a final volume of 100 μL. The mixture was incubated at 37 °C for 2 h. Samples were then mixed with loading buffer without dithiothreitol (DTT) or heating and immediately subjected to separation by sodium dodecyl sulfate−polyacrylamide gel electrophoresis (SDS−PAGE), transferred onto poly(vinylidene difluoride) membranes (PVDF), and immuno-blotted with the site-directed polyclonal antibody against the C-terminus of LacY (pAb 1043), as previously described.²⁰ The fraction of cross-linked protein was obtained by normalizing to the density of the cross-linked LacY E215Q in PE-containing cells, for each Cys pair assessed.

Circular Dichroism (CD) Spectropolarimetry.

Circular dichroism experiments were conducted on a model J-815 spectropolarimeter (Jasco) in a thermostated cell holder at 20 °C. Proteins were diluted in degassed 50 mM Tris-HCl (pH 7.5)/100 mM NaCl/0.05% DDM buffer 1 h before measurements. The scans were recorded using a bandwidth of 2 nm and an integration time of 1 s. For far-ultraviolet (far-UV) measurements, each spectrum was the average of 40 scans, with a scan rate of 100 nm/min, and with proteins at concentrations of 1 μM. The spectra were corrected for the blank, smoothed using the FFT filter (Jasco Software), and treated as previously described.³²

Substrate Binding as Assessed by Fluorescence Measurements.

The fluorescence measurements were conducted using a QuantaMaster model QM3-SS (Photon Technology International) fluorescence spectrometer. Data were collected and analyzed using Felix 32 software. All measurements were conducted in degassed 50 mM Tris-HCl (pH 7.5)/100 mM NaCl/0.05% DDM buffer. Steady state measurements were carried out in 1 cm × 1 cm cuvettes (2 mL) with constant stirring. Emission spectra were recorded with a 1 nm bandwidth for both excitation and emission. All fluorescence changes were corrected for dilution resulting from addition of the ligand. Fluorescence changes were recorded using excitation and emission wavelengths of 295 and 330 nm, respectively, for Trp → α-NPG FRET. Binding experiments were carried out by equilibrating the protein sample with increasing concentrations of α-NPG.

Ninety-Six-Well Fluorescence-Based CPM Thermo-stability Assay.

The thermostability assay was carried out as described by Stevens and co-workers.³³ The excitation wavelength was set at 387 nm, while the emission wavelength was 463 nm. One microliter of tested LacY derivative at 4 mg/mL was added to 150 μL of 50 mM Tris-HCl (pH 7.5)/100 mM NaCl/0.05% DDM buffer in a 96-well black Nunc plate. The CPM dye at 4 mg/mL in DMSO was diluted 100-fold into 50 mM Tris-HCl (pH 7.5)/100 mM NaCl/0.05% DDM buffer. After a 5 min incubation at room temperature to allow equilibration of the protein with the buffer components, 3 μL of CPM at 40 μg/mL was added, a clear plate cover was set in place, and within 5 min of protein addition fluorescence emission was measured on a BioTeK Synergy HT Microplate Reader instrument (BioTek Group). For time-dependent protein unfolding, recordings were measured at 45 °C every 1 min for 3 h with a 30 s shaking interval between each reading. The fraction of folded protein at each time point was calculated by the quotient of raw fluorescence measured at each time point divided by the maximal fluorescence measured for the series.

Homology Threading and Tertiary Structure Analysis.

Homology threading of CscB into the LacY structure was conducted using the X-ray coordinates (PDB entry 2Y5Y) as a template on the web-based SWISSPROT modeling server. All homology threading was done without manual optimization using the default mode to eliminate human bias. Three-dimensional structures obtained by homology threading were visualized and compared using Pymol 0.97 (DeLano Scientific, LLC) and compared to the LacY structure.

RESULTS

The Q353H Mutation Increases the In Vivo Transport Activity of CscB, in a PE-Independent Manner.

Plasmid pSP72 encoding the wild type or mutant gene for E. coli CscB was transformed into PE-containing, PE-lacking, and PC-containing cells. The activity of CscB was measured by its ability to actively accumulate sucrose against a concentration gradient (proton-coupled uphill transport).

The following CscB templates were analyzed: wild type CscB (WT), cysteine-less CscB (Cys-less) in which all of the native Cys residues have been replaced by Ser residues, CscB with an additional salt bridge (SB) between TMD VII (N234D) and TMD XI (S356K), which mimics the LacY counterpart in CscB, and CscB containing the Q353H mutation with (SB-Q353H) or without (Q353H) the salt bridge (Figure 2A). As previously demonstrated,²⁶ both the initial rate and steady state level of sucrose transport are severely inhibited in PE-lacking cells compared with PE-containing cells (Figure 2B). Moreover, the presence of an additional salt bridge (SB) significantly decreases the level of uphill transport of sucrose, compared to that of WT or Cys-less CscB in both PE-containing and PE-lacking cells. Quite surprisingly, no uphill sucrose transport is observed when any CscB template is expressed in PC-containing cells lacking PE. Previous reports indicated that LacY maintains close to WT levels of uphill transport when expressed in PC-containing cells,²³ challenging the long-standing observation that LacY is inactive when assembled in PC-containing membranes. This additional disparity between LacY and CscB hints at subtle changes in their TMD organization, packing, and specific lipid interactions.

Figure 2.

Uphill transport of sucrose by PE-containing, PC-containing, and PE-lacking cells. (A) Predicted topological structure of CscB based on the threaded model presented in Figure 1B. Putative TMD−TMD charge interactions are indicated, based on similarities with LacY, as well as mutation Q353H (orange). Cytoplasmically oriented EMDs are located at the top. (B) Shown is uptake of sucrose normalized to total cell protein determined as a function of time in PE-containing, PE-lacking, and PC-containing cells expressing either CscB WT, Cys-less CscB, or Cys-less CscB with the Q353H mutation (top panels) or CscB containing the additional salt bridge (SB, N234D/S356K), with or without the Q353H mutation (bottom panels). The addition of a protonophore eliminated any detectable accumulation of sucrose (FCCP). (C) Expression levels of the various CscB templates analyzed in PE-containing, PE-lacking, and PC-containing cells. After isolation of the membrane fraction, equal amounts of membrane protein (10 μg) were subjected to SDS−PAGE, followed by Western blot analysis using the LacY dodecapeptide-specific antibody. In all cases, the experiments were repeated four times, and the data represent mean values ± SD. Statistical analysis by two-way ANOVA with a Tukey multiple-comparison test. ***P <0.001.

Introducing the Q353H mutation led to significant increases in the uphill transport activity of all CscB templates in cells containing PE, which validated previous observations.²⁹ Unexpectedly, the Q353H mutation also significantly increased the level of sucrose uphill transport in PE-lacking cells, indicating the possibility of bypassing the requirement for PE. Even the low rate in the SB mutant was increased by the Q353H mutation, independent of the presence of PE. Transport was also sensitive to the addition of a protonophore (FCCP), verifying that uphill transport was occurring in both cell types. In cells in which PE was replaced with PC, no uphill sucrose transport was observed, independent of the presence of an additional SB or the Q353H mutation. We verified that these changes in transport activities were not due to variations in the expression and localization of CscB by quantifying CscB in the lipid-altered E. coli cells following inner membrane isolation, SDS−PAGE, and Western blotting analysis of CscB (Figure 2C).

Transposing the Q353H Mutation to LacY.

Next, we asked whether the effects of the Q353H mutation in CscB could be recapitulated in LacY. Unlike CscB, both the transport activity and the topological orientation of LacY are sensitive to PE. We therefore needed to use a LacY template that exhibits a correct topological orientation in PE-lacking cells to test for the rescue of uphill transport activity. The correct topological orientation in PE-lacking cells can be achieved through a single-point mutation of a negatively charged residue (E215Q) located in EMD C6.¹⁴ Using the H205C/E215Q template, which retains a partial uphill transport activity in PE-lacking cells, we then conducted a small screening of His replacement mutations along TMD XI of LacY (Figure 3A). On the basis of the sequence alignment of LacY and CscB (Figure 1C), non-essential residues in the 352−359 region of TMD XI were replaced systematically with His and tested for uphill transport activity improvements. Uphill transport of TMG by the various LacY H205C/E215Q templates was tested in both PE-containing and PE-lacking cells (Figure 3B), followed by validation of protein expression levels in the membrane (Figure 3C) and correct topological orientation in PE-lacking cells (Figure 3D) using a single-cysteine mutant of LacY (in an otherwise cysteine-less protein) with the cysteine strategically located in EMD C6 (H205C). Using SCAM,³⁴ we determined that H205C is exposed to the cytoplasm of PE-lacking cells, indicative of its correct orientation (Figure 3D). We found that, in the presence of PE, any mutations occurring in the vicinity of K358 lead to a complete loss of uphill transport activity (Figure 3B), indicating possible perturbations in the proton wire network. One mutant exhibited a significantly increased level of active transport in PE-lacking cells [Q359H (Figure 3B)], compared to the LacY H205C/E215Q template. Thus, introducing a His residue near the linker region between the two α-helical parts of TMD XI may improve the interaction arrangements inside the proton wire network in the absence of PE in the membrane, leading to a more efficient deprotonation of the permease on the inner surface of the membrane after dissociation of the sugar, and ultimately rescue of lactose/proton symport activity. While no complete loss of activity was observed, significant decreases occurred for all other mutations in PE-lacking cells. Altogether, these results indicate that the arrangement of the proton wire in LacY can be altered both by single-point protein mutations, as previously established,¹ and by changes in the membrane lipid environment.

Figure 3.

Effects of His replacements in TMD XI on the transport activity of LacY expressed in PE-containing and PE-lacking cells. (A) Topological structure of LacY indicating the residues involves in TMD−TMD charge interactions, as well as the position of the Q359H mutation (orange). Cytoplasmically oriented EMDs are located at the top. The magenta star in EMD C6 indicates the position of the Cys replacement (H205C) used for the SCAM assay. (B) Uptake of TMG normalized to total cell protein determined as a function of time in PE-containing and PE-lacking cells. The effects of His replacement in TMD XI are assessed in LacY H205C/E215Q. (C) Expression levels of the various CscB templates analyzed in PE-containing and PE-lacking cells. After isolation of the membrane fraction, equal amounts of membrane protein (10 μg) were subjected to SDS−PAGE, followed by Western blot analysis using the LacY dodecapeptide-specific antibody. (D) Determination of the orientation of EMD C6 of LacY with various His replacements in TMD XI and containing a single-cysteine replacement (H205C) using SCAM. Whole cells were labeled with MPB either before (−) or during (+) cell disruption by sonication, and the membrane extracts were subjected to immunoprecipitation, SDS−PAGE, and staining as described in Materials and Methods. In all cases, the experiments were repeated four times, and the data represent mean values ± SD. Statistical analysis by two-way ANOVA with a Tukey multiple-comparison test. **P < 0.01; ***P < 0.001.

Assessing the Consequences of the Q359H Mutation on LacY Structure, Stability, and Counterflow Activity.

To test whether the Q359H mutation had unforeseen effects on the structure and stability of LacY H205C/E215Q, we first characterized the secondary structure of DDM-solubilized LacY purified from PE-containing and PE-lacking cells using CD. We analyzed LacY templates containing a single cysteine replacement (H205C) and a charge mutation in EMD C6 (E215Q), with and without the Q359H mutation in TMD XI. In all cases, DDM-solubilized LacY exhibits typical features found in α-helix-containing proteins with two minima at 208 and 222 nm (Figure 4A). By monitoring the signal intensity at 208 and 222 nm (Figure 4B), we estimated that there was no significant change in the helicity content of LacY H205C/E215Q, independent of its cell origin and the presence of the Q359H mutation, indicating that the changes in LacY uphill transport activity are not due to any major structural alterations in its secondary structure.

Figure 4.

Effects of the Q359H mutation on LacY structure, stability, and sugar binding ability. Analysis of LacY secondary structure arrangement in DDM micelles. (A) Far-UV circular dichroism spectra and (B) mean residue ellipticities at 222 and 208 nm for LacY H205C/E215Q, with or without the Q359H mutation, isolated from PE-containing (+PE) and PE-lacking (−PE) cells. (C) Representative unfolding curves at 45 °C for 120 min for LacY H205C/E215Q, with or without the Q359H mutation, purified from PE-containing and PE-lacking cells and solubilized in DDM. (D) Binding of α-NPG to LacY H205C/E215Q, with or without the Q359H mutation, as detected by Trp → α-NPG FRET. Quenching of Trp fluorescence at different concentrations of α-NPG is depicted. Solid lines represent receptor binding − saturation (one site − total binding) fits of the data. In all cases, the experiments were repeated three times, and the data represent mean values ± SD. Statistical analysis by two-way ANOVA with a Tukey multiple-comparison test. *P < 0.05; **P < 0.01; ***P < 0.001. ns, nonsignificant.

To further test the impact of the Q359H mutation on LacY, we employed a fluorescent thermal denaturation assay using the thiol-specific dye N-[4-(7-diethylamino-4-methyl-3-coumarinyl)phenyl]maleimide (CPM) where the stability of LacY H205C/E215Q solubilized in DDM was tested as a function of time at 45 °C (Figure 4C). The results indicate that LacY H205C/E215Q purified from PE-containing cells is more stable than LacY H205C/E215Q purified from PE-lacking cells, despite the use of a mutant (H205C/E215Q) that exhibits a native topology independent of PE (see Figure 3D). However, we observed that the Q359H mutation makes LacY more stable in the absence of PE, while it appears to decrease the stability of LacY in the presence of PE, pointing toward possible rearrangements in the complex interaction network of LacY’s TMDs.

The Q359H mutation (located on TMD XI) is occurring in a region of LacY where fewer sugar interactions are observed. Nevertheless, TMDs VII (residue D237) and XI (residue K358), which are symmetrically related to TMDs I and V, respectively, may be involved in substrate binding, prompting us to test whether the Q359H mutation affects the substrate binding efficiency of LacY. We therefore tested the ability of DDM-solubilized LacY H205C/E215Q to bind the relatively high-affinity lactose analogue 4-nitrophenyl-β-d-galactopyranoside (α-NPG). To do so, we adapted a method developed by the Kaback group using a LacY template carrying a single Trp residue (W151) in the sugar-binding site.³⁵ We conducted fluorescent measurements on DDM-solubilized LacY containing the H205C and E215Q mutations and all six native Trp residues (including W151) purified from PE-containing and PE-lacking cells, with or without the Q359H mutation. Addition of α-NPG, which is the acceptor in Trp → α-NPG FRET, results in a progressive and biphasic quenching of W151 as a function of α-NPG concentration (Figure 4D). While LacY H205C/E215Q purified from PE-containing and PE-lacking cells exhibits biphasic quenching of W151, weaker quenching for the PE-lacking LacY is observed above 25 μM α-NPG, indicative of a lower α-NPG binding efficiency in the absence of PE. Despite this difference, the presence of the Q359H mutation does not lead to significant changes in the binding affinity of LacY H205C/E215Q for α-NPG for a given lipid composition, validating that the changes in uphill transport activity previously observed (Figure 3B) do not stem from a poorer ability to bind TMG. Altogether, these results indicate that the Q359H mutation does not lead to drastic alterations of the sugar-binding site of LacY H205C/E215Q, thus suggesting a possible impact on the proton wire network of interactions.

Although LacY must be protonated prior to galactopyranoside binding,^9,10 which represents a part of the coupling mechanism, it is difficult to study the mechanism of proton translocation. However, in addition to Δμ_H⁺-driven active transport, LacY catalyzes other modes of translocation that are important for studying coupled proton transport. Because individual steps in the overall transport cycle cannot be delineated by studying Δμ_H⁺-driven active transport, carrier-mediated efflux down a concentration gradient, equilibrium exchange, and entrance counterflow are used to probe the mechanism.^2,36 Equilibrium exchange and counterflow represent alternating routes of access of the galactopyranoside-binding site to alternative sides of the membrane. The efficiency of the entrance counterflow phenomenon, in kinetic terms, is due in part to the frequency with which LacY returns from the outer to the inner surface of the membrane in the loaded versus unloaded form. To further test the impact of the Q359H mutation on the properties of LacY, we analyzed the effect of pH on lactose counterflow. To do so, intact phospholipid-altered E. coli cells expressing a given LacY H205C/E215Q template were first loaded with 0.5 mM lactose and subsequently diluted 100-fold into the same buffer containing [¹⁴C]lactose at the same concentration. Experiments were conducted at five different pH values for LacY H205C/E215Q (with or without the Q359H mutation) expressed in PE-containing and PE-lacking cells (Figure 5).

Figure 5.

Effect of PE on the pH dependence of lactose counterflow at saturating external lactose concentrations. PE-containing and PE-lacking cells expressing LacY H205C/E215Q, with or without the Q359H mutation, were equilibrated with 0.5 mM lactose as described in Materials and Methods, and aliquots were diluted 100-fold into 100 mM HEPES and 50 mM MgCl₂ containing 0.5 mM [¹⁴C]lactose. Counterflow was assayed at given times at 25 °C in buffer at various pH values, ranging from 5.5 to 9.5. The inset presents the analysis of LacY expression levels following membrane isolation, SDS−PAGE separation, and Western blotting analysis of equal amounts of membrane protein (10 μg) using the LacY dodecapeptide-specific antibody. In all cases, the experiments were repeated four times, and the data represent mean values ± SD.

In the counterflow reaction, LacY binds a radiolabeled lactose molecule on the external side of the membrane and exchanges that lactose molecule for an unlabeled lactose on the internal side of the membrane. The initial 1:1 exchange of the internal unlabeled substrate for the external radiolabeled lactose leads to a rapid increase in internal radioactivity until all of the internal unlabeled lactose is exchanged for external radiolabeled lactose. At this point, the increase in internal radioactivity stops, and the internal radioactivity decreases with time as the concentrations equilibrate, thereby causing the typical “overshoot” profile that is observed (Figure 5). Typical biphasic counterflow kinetics are indeed observed for LacY H205C/E215Q expressed in PE-containing cells, where a rapid transient uptake of the radioactive substrate is observed with a peak level close to 0.5 mM lactose (indicative of 1:1 coupling between the efflux of unlabeled lactose and the influx of radioactive lactose) and relatively independent of the pH. It is then followed by a decay in [¹⁴C]lactose content, corresponding to the pH-dependent lactose efflux. In PE-lacking cells, lactose counterflow occurs at lower rates and to a lesser extent (0.2 mM). The Q359H mutation leads to improved counterflow activity in PE-lacking cells (essentially the same rate, but the internal concentration of [¹⁴C]lactose is maintained at higher levels for a markedly prolonged period relative to that of LacY H205C/E215Q in PE-lacking cells). However, the Q359H mutation results in the nearly complete abolishment of the counterflow activity in PE-containing cells in LacY H205C/E215Q, consistent with a marked defect in active transport (Figure 3B). We verified that these changes in lactose counterflow were not due to variations in the expression and localization of LacY by quantifying LacY in the lipid-altered E. coli following inner membrane isolation, SDS−PAGE, and Western blotting analysis of CscB (inset of Figure 5). Altogether, these results indicate that the absence of PE leads to a decreased efficiency in counterflow, which can be partially compensated by the Q359H mutation. Because this same mutation leads to a loss of counterflow in PE-containing cells, our results are indicative of lipid-dependent and Q359H-dependent changes in the conformational flexibility of LacY.^37,38

We then asked whether the transport-rescuing properties of the Q359H mutation observed in PE-lacking cells could be translated to other lipid compositions. In addition to the previously described complete loss of activity in PE-containing cells (Figure 6A) and rescue of activity in PE-lacking cells (Figure 6B), our results indicate that the Q359H mutation also rescues uphill transport for LacY H205C/E215Q in glycolipid-containing cells [MGDG (Figure 6D) and DGDG (Figure 6E)] but significantly impairs transport activity in PC-containing cells (Figure 6C). The same effects were observed for LacY H205C/Q359H expressed in the various phospholipid-altered strains (Figure 6, gray traces), except for PE-lacking cells in which no uphill transport occurs, validating the requirement for LacY native topology for transport rescue to occur. Finally, we verified that these changes in transport activities were not due to variations in protein expression by quantifying LacY in the various lipid-altered E. coli following inner membrane isolation, SDS−PAGE, and Western blotting analysis (Figure 6F).

Figure 6.

Effects of membrane lipid composition on the rescue of LacY uphill transport by the Q359H mutation. LacY uphill transport activity was assessed for various LacY templates expressed in (A) PE-containing cells, (B) PE-lacking cells, and cells in which PE is replaced by (C) PC, (D) MGDG, or (E) DGDG. Uptake of TMG normalized to total cell protein determined as a function of time for LacY H205C/E125Q (black), LacY H205C/E125Q/Q359H (red), and LacY H205C/Q359H (gray) is shown. (F) Expression levels of the various LacY templates analyzed in phospholipid-altered cells. After isolation of the membrane fraction, equal amounts of membrane protein (10 μg) were subjected to SDS−PAGE, followed by Western blot analysis using the LacY dodecapeptide-specific antibody. In all cases, the experiments were repeated three times, and the data represent mean values ± SD.

The Q359H Mutation Leads to Lipid-Dependent Shifts in the Helix Packing of LacY.

To assess the changes in LacY conformational flexibility, we used a combination of site-directed TEV proteolysis and cross-linking with homobifunctional thiol-reactive reagents. First, Cys-less LacY E215Q templates, with or without the Q359H mutation, were engineered to contain the TEV protease cleavage sequence (ENLYFQG) in the hydrophilic EMD P9 connecting TMDs IX and X, and two newly added Cys residues, as noted below, on either side of the cleavage site. The proximity between Cys residues was then assessed in situ by chemical cross-linking using a homobifunctional reagent with a known linker length (MTS-1-MTS, 4.2 Å). In RSO membrane vesicles prepared from E. coli phospholipid-altered strains expressing these LacY mutants and treated with TEV protease, followed by SDS−PAGE separation and Western blotting analysis, the C-terminal fragment migrates at an apparent molecular weight of ∼11 kDa (lower band) while the uncleaved LacY migrates at an apparent molecular weight of ∼33 kDa (upper band). Paired Cys residues were engineered to assess the proximity among several TMDs in various membrane lipid environments, with or without the Q359H mutation. Specifically, the D237C−K358C, D240C−H322C, and E269C−H322C pairs were used to monitor TMD VII−TMD XI, TMD VII−TMD X, and TMD VIII−TMD X distances, respectively. Cross-linking experiments were conducted as previously described.³⁹ When cross-linking occurs, a band corresponding to intact LacY migrating at ∼33 kDa is observed after proteolytic treatment.

As depicted in Figure 7A, LacY E215Q containing the D237C and K358C mutations is cross-linked by MTS-1-MTS in PE-containing membranes, indicating that the distance between TMDs VII and XI is ∼3−6 Å as previously established.³⁹ A similar level of cross-linking (histograms in Figure 7A) is observed in PC-containing cells, hinting at maintenance of the TMD−TMD distance. However, an increased cross-linking efficiency is detected in PE-lacking and MGDG-containing cells, indicative of the closer proximity of residues D237C and K358C, while a decreased cross-linking efficiency is observed in DGDG-containing cells, indicative of an increased distance between these Cys residues. Decreased cross-linking efficiencies are observed between residues D240C and H322C in PE-lacking and DGDG-containing cells (Figure 7B), indicating greater separation between TMDs VII and X; an increased cross-linking efficiency is observed in PC-containing cells, and no changes are observed in MGDG-containing cells. Increased cross-linking efficiencies are observed between residues E269C and H322C in all phospholipid-altered strains, compared to PE-containing cells (Figure 7C), illustrating that TMDs VII and X are closer to each other under these conditions.

Figure 7.

Analysis of the effects of the membrane lipid environment on LacY conformational flexibility. The cross-linking efficiency between Cys pairs (A) D237C and K358C, (B) D240C and H322C, or (C) E269C and H322C in LacY E215Q is depicted, where increased levels of cross-linking indicate increased proximity between helices and lower levels of cross-linking indicate increased distances between helices. Histograms depict cross-linking efficiency, which was estimated after densitometric quantification of the upper, cross-linked bands. Chemical cross-linking using MTS-1-MTS was conducted on RSO membrane vesicles containing LacY E215Q with paired cysteine replacements. RSO membrane vesicles were incubated with 50 μM MTS-1-MTS for 1 h at 4 °C. Reactions were terminated by addition of NEM to a final concentration of 5 mM, followed by cleavage with TEV protease for 2 h, at 37 °C. Samples were then subjected to SDS−PAGE and Western blotting analysis with the anti-C-terminal antibody. In all cases, the experiments were repeated three times, and the data represent mean values ± SD. Statistical analysis by two-way ANOVA with a Tukey multiple-comparison test. *P < 0.05; ***P < 0.001. ns, nonsignificant.

In the presence of the Q359H mutation, opposite changes in cross-linking efficiencies are observed in PE-containing cells for the TMD VII−TMD X and TMD VIII−TMD X pairs, with residues D240C and H322C moving closer to each other and E269C and H322C moving farther apart. These changes in distances between key TMDs for the proton translocation pathway may explain the loss of uphill transport and counterflow activities observed for LacY H205C/E215Q/Q359H expressed in PE-containing cells. Changes in distances between TMDs in phospholipid-altered strains where active transport was partially rescued are also observed (PE-lacking cells and MGDG- and DGDG-containing cells), hinting at a possible realignment of these TMDs that could increase the conformational flexibility of LacY in the absence of PE. Minimal changes in TMD distances are observed in PC-containing cells upon introduction of the Q359H mutation with only residues D237C and K358C moving closer to each other, possibly reflecting the impairing effect of the Q359H mutation on LacY transport activity.

The Q359H Mutation Leads to Lipid-Dependent Alterations in the Folding of the 4B1 Epitope in EMD P7.

The native structure of the periplasmic domain P7 connecting TMDs VII and VIII contains a continuous epitope that is recognized by the conformation-specific mAb 4B1. The proper folding of this epitope is correlated with uphill transport of lactose in intact cells,⁷ making the recognition by mAb 4B1 a valid indicator of the proper topological and structural organization of LacY in the vicinity of domain P7. We therefore assessed the conformation of the 4B1 epitope in intact cells using mAb 4B1 immunoprecipitation of LacY H205C/E215Q, as previously described.²⁰

Representative Coomassie-stained gels after immunoprecipitation and SDS−PAGE separation are presented for the various phospholipid-altered E. coli strains used. As a control, pAb 1043 (which detects LacY irrespective of orientation or functionality) was used to assess immunoprecipitation of total LacY regardless of 4B1 epitope folding. The histogram presented in Figure 8 depicts the proportion of LacY with a properly folded EMD P7, as assessed by mAb 4B1/pAb 1043 immunoprecipitation of LacY. Binding of mAb 4B1 to LacY H205C/E215Q assembled in the membrane of phospholipid-altered E. coli strains exhibits significant differences compared to PE-containing membranes (Figure 8), reminiscent of our previously published work on LacY reconstituted in proteoliposomes made of various lipid compositions.²⁰ PE-and PC-containing cells support the proper folding of the 4B1 epitope, while glycolipid-containing cells support only partial folding. Despite exhibiting the correct topological orientation and partial transport activity, LacY assembled in PE-lacking cells exhibits a small amount of 4B1 epitope folding.

Figure 8.

Effect of lipids on the folding of the 4B1 epitope of LacY. LacY H205C/E215Q expressed in phospholipid-altered strains was immunoprecipitated using either conformation-sensitive mAb 4B1 or pAb 1043, each previously cross-linked to Protein A/G-agarose beads. After SDS−PAGE analysis, visualization was conducted by Coomassie Blue staining. Histograms depict band intensity ratios for the different phospholipid-altered strains expressing LacY H205C/E215Q (black bars) or LacY H205C/E215Q/Q359H (gray bars). In all cases, the experiments were repeated three times, and the data represent mean values ± SD. Statistical analysis by two-way ANOVA with a Tukey multiple-comparison test. **P < 0.01; ***P < 0.001. ns, nonsignificant.

When the Q359H mutation is engineered into LacY H205C/E215Q, significant increases in the level of 4B1 folding are observed in PE-lacking and glycolipid-containing cells, correlating with improvement in active transport and counterflow efficiencies. In contrast, the Q359H mutation leads to significant decreases in the level of 4B1 folding in PE- and PC-containing cells, in agreement with their impaired active transport and counterflow activities. Altogether, these results indicate that the Q359H mutation leads to significant structural rearrangements in the interaction network of LacY TMDs, which in turn lead to alterations in LacY function. The ultimate outcome of these changes is dependent on the lipid environment in which LacY is assembled, shedding new light on the previously overlooked role of membrane phospholipids on the mechanism of sugar/proton symport by LacY.

DISCUSSION

To further our understanding of how lipid−protein interactions govern protein structure, high-resolution structures of membrane proteins as a function of lipid environment must be obtained. However, solubilized proteins must first be verified to reflect the structure and function of the protein in the cell membrane. Atomic-resolution structures of membrane proteins provide an invaluable amount of information, but the importance of lipid−protein interactions in membrane protein structure and function remains poorly tractable using these methods.

Members of the MFS present with both great functional diversity and well-conserved architecture, indicating that some mechanistic features may be conserved between members of the family.⁴⁰ Indeed, members of the MFS are highly dynamic proteins undergoing significant structural changes to transport their substrate across the membrane. LacY is considered as a paradigm for membrane transport proteins as it has been extensively explored by a number of molecular biological, biochemical, and biophysical approaches. It has been clearly established that deprotonation of E325 is coupled with significant conformational changes in LacY, which in turn lead to the transition between inward- and outward-facing conformations.^41,42 Molecular dynamics simulations have highlighted the importance of specific lipid−LacY interactions in the process, helping in bridging the information gap between static crystal structures and LacY conformational flexibility.^43–46

Comparison between “optimal” (palmitoyl-oleyl-PE-containing, POPE) and “non-optimal” (dimyristoyl-PC-containing, DMPC) membrane environments, i.e., lipids that either support (POPE) or do not support (DMPC) uphill transport of LacY,⁴⁶ led to the discovery that LacY conformational flexibility depends on both E325 deprotonation and the membrane lipid environment. Indeed, changes in several residue−residue interactions observed in POPE-containing membranes are not occurring in DMPC-containing membranes. For example, the shortening of the distance between H322 and E269 upon E325 deprotonation is abrogated in the presence of DMPC. Altogether, these simulations⁴⁶ indicate that the large conformational changes observed upon E325 deprotonation of LacY embedded in POPE membranes do not occur in DMPC membranes. Further analyses identified that PC headgroups did not interact with LacY, while several residues displayed direct interactions with the PE headgroup, including two residues located at the periplasmic end of TMD XI (N371 and E374). These MDSs suggest that LacY engages in a diverse array of connections with its close lipid environment.⁴⁶

These predictions are very reminiscent of the in vitro findings from the Borrell Hernandez group, which demonstrated by FRET between pyrene-labeled lipids and LacY containing a single Trp residue that PE preferentially associates with LacY with the exclusion of PG and CL. However, PG preferentially interacts with LacY when DOPC (dioleyl-PC) or POPC (palmitoyl-oleyl-PC) is present in the membrane, validating the absence of direct interactions between LacY and PC headgroups.⁴⁷

Our results provide the first in vivo insights into a lipid-dependent control of the LacY conformational equilibrium. In the absence of PE, but in the presence of a native LacY topology, the structural arrangement of LacY TMDs inside the membrane is altered, resulting in several shifts and changes in key residue interactions, which in turn lead to an imbalance in the conformational switch involved in substrate transport. The conformation of EMD P7, as indicated by recognition with mAb 4B1, is directly correlated with the ability of LacY to carry out uphill transport that requires protonation of E325. Binding of mAb 4B1 to LacY inhibits uphill transport in spheroplasts and in proteoliposomes⁴⁸ concomitant with conformational changes in several TMDs.⁴⁹ Therefore, structural changes within EMD P7 induced by binding of mAb 4B1, and by extrapolation folding defects in EMD P7 due to a lack of PE, result in a conformational change that decreases the pK_a of an essential acidic residue at position 325, which is involved in proton-coupled uphill transport. We have determined that EMD P7 is partially to properly folded when net neutral lipids are present but is misfolded in their absence. Improper folding of EMD P7 results in a shift in the tilt angles of TMDs^8,49 with a loss of uphill transport activity. These alterations can be partially corrected by a single-point mutation in LacY (Q359H) located near the proton wire network, which somewhat realigns the various TMDs and rescues uphill transport activity.

This single-point mutation was originally found in CscB (Q353H).²⁹ Although very similar to LacY, the charge pair between D237 (TMD VII) and K358 (TMD XI) is not conserved in CscB. If mutations N234D and S356K are simultaneously introduced into CscB to recreate the salt bridge present in LacY, a significant decrease in active sucrose transport activity is observed (Figure 2B).²⁶ Most interesting is that introduction of Q353H into TMD XI of CscB not only improved uphill transport function in PE-containing cells but also suppressed the lack of uphill transport activity in PE-lacking cells. This appears to be due to the introduction of a positively charged proton acceptor (histidine) in the proton wire, which unlike glutamate or aspartate is insensitive to the absence of PE.

Due to the high degree of conservation observed between transporters in the proton translocation network,⁵⁰ we anticipate that these observations could be translated to other sugar transporters, such as RafB and MelB. Such flexibility and/or capability of a given transporter to accommodate the local lipid environment and optimize its function through a minimal degree of mutation may have contributed to the spread of MFS members across species and organisms, which exhibit different membrane lipid compositions. Further studies of both the impact of phospholipid and fatty acid composition could be pursued both in vivo and in vitro but would require a complete characterization of the fatty acid diversity in each of the major phospholipid classes.

ABBREVIATIONS

ANOVA

analysis of variance

BCA

microbicinchoninic acid

CL

cardiolipin

CPM

N-[4-(7-diethylamino-4-methyl-3-coumarinyl)phenyl]maleimide

CscB

sucrose permease

CT

C-terminal

DDM

n-dodecyl-β-d-maltoside

DGDG

diglucosyl diacylglycerol

DMPC

dimyristoyl-phosphatidylcholine

DOPC

dioleyl-phosphatidylcholine

DTT

dithiothreitol

EMD

extramembrane domain

HRP

horseradish peroxidase

FCCP

carbonyl cyanide-p-trifluoromethoxyphenylhydrazone

IPTG

isopropyl-β-thiogalactoside

LacY

lactose permease

mAb 4B1

monoclonal antibody against the LacY 4B1 epitope

MDS

molecular dynamics simulation

MelB

melibiose permease

MGDG

monoglucosyl diacylglycerol

MPB

3-(N-maleimidopropionyl) biocytin

MTS-1-MTS

1,1-methanediyl bismethanethiosulfonate

NEM

N-ethylmaleimide

NPG

4-nitrophenyl-β-d-galactopyranoside

NT

N-terminal

pAb 1043

polyclonal antibody directed against the C-terminal dodecapeptide of LacY

PC

phosphatidylcholine

PDB

Protein Data Bank

POPC

palmitoyl-oleyl-phosphatidylcholine

PE

phosphatidylethanolamine

PG

phosphatidylglycerol

POPE

palmitoyloleyl-phosphatidylethanolamine

PS

phosphatidylserine

RafB

raffinose permease

SCAM

substituted cysteine accessibility method for determining TMD orientation

SD

standard deviation

TEV

tobacco etch virus

TMD

transmembrane domain

TMG

methyl-thio-β-d-galactopyranoside.