The Escherichia coli multidrug transporter EmrE is a dimer in the detergent-solubilised state

EmrE is a multidrug transporter that utilises the proton gradient across bacterial cell membranes to pump hydrophobic cationic toxins out of the cell. The structure of EmrE is very unusual, because it is an asymmetric homodimer containing eight alpha-helices, six of which form the substrate-binding chamber and translocation pathway. Despite this structural information, the precise oligomeric order of EmrE in both the detergent-solubilised state and in vivo is unclear, although it must contain an even number of subunits to satisfy substrate-binding data. We have studied the oligomeric state of EmrE, purified in a functional form in dodecylmaltoside, by high-resolution size-exclusion chromatography (hrSEC) and by analytical ultracentrifugation. The data from equilibrium analytical ultracentrifugation were analysed using a measured density increment for the EmrE-lipid-detergent complex, which showed that the purified EmrE was predominantly a dimer. This value was consistent with the apparent mass for the EmrE-lipid-detergent complex (137 kDa) determined by hrSEC. EmrE was purified under different conditions using minimal concentrations of dodecylmaltoside, which would have maintained the structure of any putative higher oligomeric states: this EmrE preparation had an apparent mass of 206 kDa by hrSEC and equilibrium analytical ultracentrifugation showed unequivocally that EmrE was a dimer, although it was associated with a much larger mass of phospholipid. In addition, the effect of the substrate tetraphenylphosphonium on the oligomeric state was also analysed for both preparations of EmrE; velocity analytical ultracentrifugation showed that the substrate had no effect on the oligomeric state. Therefore, in the detergent dodecylmaltoside and under conditions where the protein is fully competent for substrate binding, EmrE is dimeric and there is no evidence from our data to suggest higher oligomeric states. These observations are discussed in relation to the recently published structures of EmrE from two- and three-dimensional crystals.

New insights into the structure and oligomeric state of the bacterial multidrug transporter EmrE: an unusual asymmetric homo-dimer

EmrE is a small multidrug transporter that contains 110 amino acid residues that form four transmembrane alpha-helices. The three-dimensional structure of EmrE has been determined from two-dimensional crystals by electron cryo-microscopy. EmrE is an asymmetric homo-dimer with one substrate molecule bound in a chamber accessible laterally from one leaflet of the lipid bilayer. Evidence from substrate binding analyses and analytical ultracentrifugation of detergent-solubilised EmrE shows that the minimum functional unit for substrate binding is a dimer. However, it is possible that EmrE exists as a tetramer in vivo and plausible models are suggested based upon analyses of two-dimensional crystals.

Overexpression of mammalian integral membrane proteins for structural studies

Recent successes in the determination of atomic resolution structures of integral membrane proteins have relied on purifying the proteins from abundant natural sources. In contrast, the majority of mammalian receptors, ion channels and transporters need to be overexpressed to obtain sufficient material for structural studies. This has often proved to be very difficult. Overexpression studies on a wide range of mammalian membrane proteins have shown that a few can be expressed functionally in bacteria, but many others require an insect or mammalian cell host for activity or high level expression. The serotonin transporter, which has been expressed in all the major hosts available, is a good example that has given insights into the problem of overexpressing mammalian membrane proteins for structural studies.

Small is mighty: EmrE, a multidrug transporter as an experimental paradigm

EmrE is a multidrug transporter from Escherichia coli that functions as a homooligomer and is unique in its small size. In each monomer there are four tightly packed transmembrane segments and one membrane-embedded charged residue. This residue provides the basis for the coupling mechanism as part of a binding site "time shared" by substrates and protons.

Biophysical characterization of the cocaine binding pocket in the serotonin transporter using a fluorescent cocaine analogue as a molecular reporter

To explore the biophysical properties of the binding site for cocaine and related compounds in the serotonin transporter SERT, a high affinity cocaine analogue (3beta-(4-methylphenyl)tropane-2beta-carboxylic acid N-(N-methyl-N-(4-nitrobenzo-2-oxa-1,3-diazol-7-yl)ethanolamine ester hydrochloride (RTI-233); K(I) = 14 nm) that contained the environmentally sensitive fluorescent moiety 7-nitrobenzo-2-oxa-1,3-diazole (NBD) was synthesized. Specific binding of RTI-233 to the rat serotonin transporter, purified from Sf-9 insect cells, was demonstrated by the competitive inhibition of fluorescence using excess serotonin, citalopram, or RTI-55 (2beta-carbomethoxy-3beta-(4-iodophenyl)tropane). Moreover, specific binding was evidenced by measurement of steady-state fluorescence anisotropy, showing constrained mobility of bound RTI-233 relative to RTI-233 free in solution. The fluorescence of bound RTI-233 displayed an emission maximum (lambda(max)) of 532 nm, corresponding to a 4-nm blue shift as compared with the lambda(max) of RTI-233 in aqueous solution and corresponding to the lambda(max) of RTI-233 in 80% dioxane. Collisional quenching experiments revealed that the aqueous quencher potassium iodide was able to quench the fluorescence of RTI-233 in the binding pocket (K(SV =) 1.7 m(-)(1)), although not to the same extent as free RTI-233 (K(SV =) 7.2 m(-)(1)). Conversely, the hydrophobic quencher 2,2,6,6-tetramethylpiperidine-N-oxyl (TEMPO) quenched the fluorescence of bound RTI-233 more efficiently than free RTI-233. These data are consistent with a highly hydrophobic microenvironment in the binding pocket for cocaine-like uptake inhibitors. However, in contrast to what has been observed for small-molecule binding sites in, for example, G protein-coupled receptors, the bound cocaine analogue was still accessible for aqueous quenching and, thus, partially exposed to solvent.

Molecular chaperones stimulate the functional expression of the cocaine-sensitive serotonin transporter

The serotonin transporter (SERT) is an N-glycosylated integral membrane protein that is predicted to contain 12 transmembrane regions. SERT is the major binding site in the brain for antidepressant drugs, and it also binds amphetamines and cocaine. The ability of various molecular chaperones to interact with a tagged version of SERT (Myc-SERT) was investigated using the baculovirus expression system. Overexpression of Myc-SERT using the baculovirus system led to substantial quantities of inactive transporter, together with small amounts of fully active and, therefore, correctly folded molecules. The high levels of inactive Myc-SERT probably arose because folding was rate-limiting due, perhaps, to insufficient molecular chaperones. Therefore, Myc-SERT was co-expressed with the endoplasmic reticulum (ER) molecular chaperones calnexin, calreticulin and immunoglobulin heavy chain binding protein (BiP), and the foldase, ERp57. The expression of functional Myc-SERT, as determined by an inhibitor binding assay, was enhanced nearly 3-fold by co-expressing calnexin, and to a lesser degree on co-expression of calreticulin and BiP. Co-expression of ERp57 did not increase the functional expression of Myc-SERT. A physical interaction between Myc-SERT-calnexin and Myc-SERT-calreticulin was demonstrated by co-immunoprecipitation. These associations were inhibited in vivo by deoxynojirimycin, an inhibitor of N-glycan precusor trimming that is known to prevent the calnexin/calreticulin-N-glycan interaction. Functional expression of the unglycosylated SERT mutant, SERT-QQ, was also increased on co-expression of calnexin, suggesting that the interaction between calnexin and SERT is not entirely dictated by the N-glycan. SERT is the first member of the neurotransmitter transporter family whose folding has been shown to be assisted by the molecular chaperones calnexin, calreticulin, and BiP.

Heterologous expression of G-protein-coupled receptors

G-protein-coupled receptors (GPCRs) are integral membrane proteins of great pharmacological importance owing to their central role in the regulation of cellular responses to external stimuli. Heterologous expression systems have been used to explore ligand binding, G protein and effector coupling, and structural aspects of the receptors. GPCRs can be expressed in a functional form in all expression systems, but with varying degrees of success because of differences in receptor and host cell characteristics. This article will discuss aspects related to the choice and suitability of expression systems for the intended analysis of GPCR properties.

Inability to N-glycosylate the human norepinephrine transporter reduces protein stability, surface trafficking, and transport activity but not ligand recognition

The role of N-glycosylation in the expression, stability, and ligand recognition by the cocaine- and antidepressant-sensitive human norepinephrine transporter (hNET) was assessed in stably and transiently transfected cell lines. The use of hNET-specific antibodies and the membrane-impermeant biotinylating reagent sulfosuccinimidobiotin establishes that treatment of stably transfected LLC-PK1 cells with tunicamycin depletes surface membranes of mature hNET glycoproteins, which is consistent with a failure of less stable, nonglycosylated subunits to replenish surface compartments. To determine whether N-glycosylation plays a direct role in hNET stability, surface expression, and ligand recognition, we mutated the three hNET canonical N-glycosylation sites (hNETN184, 192, 198Q) and transiently expressed the mutant cDNA in parallel with the parental hNET construct in HeLa and COS cells. hNETN184, 192, 198Q protein exhibited increased electrophoretic mobility (approximately 46 kDa), similar to that of enzymatically N-deglycosylated hNET protein, which confirms the use of canonical sites in the second extracellular loop of the transporter. hNETN184, 192, 198Q protein in HeLa and COS extracts was reduced approximately 50% relative to hNET protein in parallel transfections, demonstrated to arise from a reduction in transporter half-life, which is consistent with the proposed role of N-glycosylation in hNET stability. Both HeLa and COS cells transfected with hNETN184, 192, 198Q exhibit a significantly greater reduction in transport activity than can be accounted for by losses in either total or surface NET protein. Furthermore, sensitivity of catecholamine transport to unlabeled substrate and antagonists was unchanged in the mutant, suggesting that residual nonglycosylated surface hNETs execute a key step in the transport cycle after ligand recognition with reduced efficiency.

Overexpression of integral membrane proteins for structural studies

Determination of the structure of integral membrane proteins is a challenging task that is essential to understand how fundamental biological processes (such as photosynthesis, respiration and solute translocation) function at the atomic level. Crystallisation of membrane proteins in 3D has led to the determination of four atomic resolution structures [photosynthetic reaction centres (Allenet al. 1987; Changet al. 1991; Deisenhofer & Michel, 1989; Ermleret al. 1994); porins (Cowanet al. 1992; Schirmeret al. 1995; Weisset al. 1991); prostaglandin H2synthase (Picotet al. 1994); light harvesting complex (McDermottet al. 1995)], and crystals of membrane proteins formed in the plane of the lipid bilayer (2D crystals) have produced two more structures [bacteriorhodopsin (Hendersonet al. 1990); light harvesting complex (Kühlbrandtet al. 1994)].

Topological analyses of the L-fucose-H+ symport protein, FucP, from Escherichia coli

The transport of L-fucose into Escherichia coli is mediated by the L-fucose-H+ symport protein (FucP). The fucP gene has been sequenced and encodes a hydrophobic protein that contains 438 amino acid residues, with a predicted M(r) of 47,773. The hydropathic profile of FucP indicates 10 to 12 hydrophobic regions that could span the membrane as alpha-helices. A 12-helix model with the N- and C-termini located in the cytoplasm was derived from the hydropathic profile and from application of the 'positive inside' rule. This model was tested using beta-lactamase fusion technology. Analyses of 62 different FucP-beta-lactamase fusions suggested that the FucP protein crosses the cytoplasmic membrane of E. coli 12 times, with the N- and C-termini in the cytoplasm. From measurements of [14C]-L-fucose uptake, it was deduced that the last putative transmembrane region must be complete for transport activity to be retained and that the four C-terminal residues were unnecessary for transport activity. Fourier transform analyses show that all the predicted helices contain a periodicity that enables hydrophobic/hydrophilic faces to be identified; these were particularly evident in putative helices 1, 3, 4, 5, 6, 10 and 11.

The effect of N-linked glycosylation on activity of the Na(+)- and Cl(-)-dependent serotonin transporter expressed using recombinant baculovirus in insect cells

The rat Na(+)- and Cl(-)-dependent serotonin transporter was expressed in Sf9 insect cells using the baculovirus system. Expression of the serotonin transporter caused the Sf9 cells to accumulate [3H]serotonin (Km 78 nM) and to bind the specific transport inhibitor [125I]RT155 (2 beta-carbomethoxy-3 beta-(4-[125I]iodophenyl)tropane) (Kd 0.22 nM). Ligand binding assays on isolated membranes showed 500,000 copies of the serotonin transporter/cell (9 pmol/mg of membrane protein). Immunoreactive bands of apparent M(r) 54,000 (unglycosylated) and 60,000 (glycosylated) were observed in Western blots of membrane proteins from infected cells. The 54-kDa band was significantly smaller than the expected M(r) of 72,500 predicted from the cDNA sequence. The 54-kDa band was shown to represent the intact serotonin transporter by expressing a recombinant serotonin transporter that contained c-Myc and FLAG epitope tags engineered at the N and C termini, respectively. Both tags were present on a membrane protein that migrated slightly slower than the previously observed 54-kDa band, consistent with the extra mass added by the tags. The tags did not affect the Kd for [125I]RT155 binding. The effect of N-linked glycosylation on ligand binding and the level of expression were studied. The expression of the serotonin transporter in tunicamycin-treated Sf9 cells resulted in low levels of ligand binding activity (0.2 pmol/mg) but unchanged Kd. Similarly, mutated serotonin transporters that contained reduced numbers of N-linked glycosylation sites had unchanged Kd for [125I]RT155 binding whether there were 2, 1, or 0 N-linked glycosylation sites present on the serotonin transporter. In contrast, Bmax was dramatically reduced; levels of expression of the unglycosylated serotonin transporter (0.4 pmol/mg) were 20-fold lower compared with levels of the fully glycosylated serotonin transporter. The Km for [3H]serotonin uptake was also unchanged. These data indicate that glycosylation is required for optimal stability of the serotonin transporter in the membrane but not for serotonin transport or ligand binding per se.

Identification of a novel sugar-H+ symport protein, FucP, for transport of L-fucose into Escherichia coli

L-Fucose (6-deoxy-L-galactose) is used as sole carbon source by many microorganisms, and its transport into Escherichia coli is mediated by an L-fucose-H+ symport activity. In order to determine the nature of a putative transporter encoded by the E. coli fucP gene and identify its protein product it was cloned downstream of the inducible T7 RNA polymerase and lambda OLPL promoters. Induction of the T7 promoter resulted in the expression of [14C]-L-fucose uptake activity and the concomitant expression of a [35S]-Met-labelled 32 kDa protein at levels too low for detection by staining with Coomassie brilliant blue or for protein sequencing. Induction of the lambda OLPL promoter caused the appearance of L-fucose-H+ symport activity and of a Coomassie brilliant blue-stained 32 kDa membrane protein expressed at high levels sufficient for identification as FucP by N-terminal protein sequencing. The FucP protein is, therefore, a sugar-H+ symporter different in amino acid sequence from any other known transporter. These and other results illustrate the general unpredictability of cloning strategies for attempting the amplified expression of membrane transport proteins.

Membrane topology of the L-rhamnose-H+ transport protein (RhaT) from enterobacteria

The L-rhamnose-H+ symporter (RhaT) is a 344-amino acid integral membrane protein, found in many Enterobacteria, which couples the uptake of the sugar L-rhamnose with the inward movement of protons. Based on its hydropathy profile and the application of von Heijne's "positive inside" rule (von Heijne, G. (1992) J. Mol. Biol. 225, 487-494), a model of the L-rhamnose-H+ symport protein (RhaT) is proposed containing 10 transmembrane helices with the NH2 and COOH termini in the periplasm. This model was tested by the creation of random beta-lactamase (Bla) fusions. The data from 33 unique, randomly generated, RhaT-Bla fusions and from 5 site-specific fusions supported the proposed topology between transmembrane helices 2-10. However, the localization of the putative first hydrophilic loop and the NH2 terminus was not possible because the beta-lactamase fusions in this region were shown to be unreliable indicators of the topology of RhaT.

Proton-linked L-rhamnose transport, and its comparison with L-fucose transport in Enterobacteriaceae

1. An alkaline pH change occurred when L-rhamnose, L-mannose or L-lyxose was added to L-rhamnose-grown energy-depleted suspensions of strains of Escherichia coli. This is diagnostic of sugar-H+ symport activity. 2. L-Rhamnose, L-mannose and L-lyxose were inducers of the sugar-H+ symport and of L-[14C]rhamnose transport activity. L-Rhamnose also induced the biochemically and genetically distinct L-fucose-H+ symport activity in strains competent for L-rhamnose metabolism. 3. Steady-state kinetic measurements showed that L-mannose and L-lyxose were competitive inhibitors (alternative substrates) for the L-rhamnose transport system, and that L-galactose and D-arabinose were competitive inhibitors (alternative substrates) for the L-fucose transport system. Additional measurements with other sugars of related structure defined the different substrate specificities of the two transport systems. 4. The relative rates of H+ symport and of sugar metabolism, and the relative values of their kinetic parameters, suggested that the physiological role of the transport activity was primarily for utilization of L-rhamnose, not for L-mannose or L-lyxose. 5. L-Rhamnose transport into subcellular vesicles of E. coli was dependent on respiration, was optimal at pH 7, and was inhibited by protonophores and ionophores. It was insensitive to N-ethylmaleimide or cytochalasin B. 6. L-Rhamnose, L-mannose and L-lyxose each elicited an alkaline pH change when added to energy-depleted suspensions of L-rhamnose-grown Salmonella typhimurium LT2, Klebsiella pneumoniae, Klebsiella aerogenes, Erwinia carotovora carotovora and Erwinia carotovora atroseptica. The relative rates of subsequent acidification varied, depending on both the organism and the sugar. L-Fucose promoted an alkaline pH change in all the L-rhamnose-induced organisms except the Erwinia species. No L-rhamnose-H+ symport occurred in any organism grown on L-fucose. 7. All these results showed that L-rhamnose transport into the micro-organisms occurred by a system different from that for L-fucose transport. Both systems are energized by the trans-membrane electrochemical gradient of protons. 8. Neither steady-state kinetic measurements nor binding-protein assays revealed the existence of a second L-rhamnose transport system in E. coli.

Mapping, cloning, expression, and sequencing of the rhaT gene, which encodes a novel L-rhamnose-H+ transport protein in Salmonella typhimurium and Escherichia coli

A L-rhamnose transport-negative strain of Escherichia coli was generated by Mu d(ApR,lac)I mutagenesis. This strain was used to isolate a clone of Salmonella typhimurium DNA that encoded L-rhamnose-H+ transport activity, the gene for which, rhaT, was sequenced. The rhaT gene was mapped on the E. coli chromosome between rhaR and sodA at 87.9 min, initially by Southern blot analysis and then by the isolation, expression, and sequencing of the rhaT gene. Both rhaT genes encoded a hydrophobic protein of 344 amino acids (91% identical) that contained 10 putative transmembrane regions. The RhaT protein represents a novel class of sugar transport protein.

Preferential expression of the complement regulatory protein decay accelerating factor at the fetomaternal interface during human pregnancy

The expression of decay accelerating factor (DAF) was investigated in human fetal and extra-fetal tissues using a panel of mAb directed against different epitopes on the DAF molecule. By immunostaining, extensive reactivity was observed on the placental trophoblast epithelium and this was confined exclusively to sites of direct contact with maternal blood and tissues at the fetomaternal interface. Within the fetus, by contrast, very little staining was observed especially on hemopoietic cell populations in the developing liver. The antibodies identified a component of 70,000 m.w., comigrating on SDS-PAGE with red cell DAF, on isolated trophoblast membranes and an apparently quantitative increase in the expression of DAF was observed during placental development. Northern analysis using a DAF cDNA clone revealed 1.5-, 1.6-, and 2.2-kb mRNA transcripts typical of DAF in mRNA prepared from whole term placentae and from purified trophoblast cells. DAF appears to be preferentially expressed at the fetomaternal interface during development and may function specifically to inhibit amplification convertases formed at this site either directly or indirectly as a result of maternal complement activation. This molecule may play an important role in protecting the semiallogeneic human conceptus from maternal C-mediated attack.

Preferential expression of the complement regulatory protein decay accelerating factor at the fetomaternal interface during human pregnancy

The expression of decay accelerating factor (DAF) was investigated in human fetal and extra-fetal tissues using a panel of mAb directed against different epitopes on the DAF molecule. By immunostaining, extensive reactivity was observed on the placental trophoblast epithelium and this was confined exclusively to sites of direct contact with maternal blood and tissues at the fetomaternal interface. Within the fetus, by contrast, very little staining was observed especially on hemopoietic cell populations in the developing liver. The antibodies identified a component of 70,000 m.w., comigrating on SDS-PAGE with red cell DAF, on isolated trophoblast membranes and an apparently quantitative increase in the expression of DAF was observed during placental development. Northern analysis using a DAF cDNA clone revealed 1.5-, 1.6-, and 2.2-kb mRNA transcripts typical of DAF in mRNA prepared from whole term placentae and from purified trophoblast cells. DAF appears to be preferentially expressed at the fetomaternal interface during development and may function specifically to inhibit amplification convertases formed at this site either directly or indirectly as a result of maternal complement activation. This molecule may play an important role in protecting the semiallogeneic human conceptus from maternal C-mediated attack.

Studies on human red-cell membrane glycophorin A and glycophorin B genes in glycophorin-deficient individuals

1. Genomic DNA derived from individuals who lack glycophorin A (GPA), glycophorin B (GPB) or both of these proteins was subjected to Southern-blot analysis using GPA and GPB cDNA probes. 2. Bands on the Southern blots were assigned to the GPA gene, GPB gene or to a putative pseudogene. 3. Genomic DNA derived from an individual of the Mk phenotype was shown to have deletions in the GPA and GPB genes. The simplest model for the results obtained is that a single deletion spans the GPA and GPB genes in the individual studied.

Studies on the defect which causes absence of decay accelerating factor (DAF) from the peripheral blood cells of an individual with the Inab phenotype

1. We have studied the peripheral blood cells of an individual with the Inab phenotype who is deficient in decay accelerating factor (DAF). 2. In contrast with the situation in paroxysmal nocturnal haemoglobinuria, membranes from peripheral blood cells of the Inab phenotype individual lack DAF, but retain the other glycosylphosphatidylinositol-linked proteins acetylcholinesterase and LFA-3. 3. Unlike normal Epstein-Barr-virus-transformed lymphoblastoid cell lines (EBV-LCL), DAF was not expressed on EBV-LCL derived from peripheral blood lymphocytes of the Inab individual. 4. No differences in the DAF gene of normal and Inab phenotype individuals could be detected by Southern blotting studies. 5. EBV-LCL derived from the Inab individual had a gross reduction in the level of DAF mRNA compared with normal EBV-LCL. 6. Our results suggest that the DAF gene in the Inab phenotype contains a mutation which affects the transcription or processing of DAF mRNA.

Isolation of cDNA clones for human erythrocyte membrane sialoglycoproteins alpha and delta

We have isolated almost full-length cDNA clones corresponding to human erythrocyte membrane sialoglycoproteins alpha (glycophorin A) and delta (glycophorin B). The predicted amino acid sequence of delta differs at two amino acid residues from the sequence determined by peptide sequencing. The sialoglycoprotein delta clone we have isolated contains an interrupting sequence within the region that gives rise to the cleaved N-terminal leader sequence for the protein and represents a product that is unlikely to be inserted into the erythrocyte membrane. Comparison of the cDNA sequences of alpha and delta shows very strong homology at the DNA level within the coding regions. The two mRNA sequences are closely related and differ by a number of clearly defined insertions and deletions.