Aminoglycosides Are Captured from both Periplasm and Cytoplasm by the AcrD Multidrug Efflux Transporter of Escherichia coli

Bacteria, plasmids, and growth conditions. E. coli KAM3 (a ΔacrB mutant of K-12 [20]) transformed with plasmid pTrcHacrD, obtained from A. Yamaguchi, was used for the expression of the RND protein AcrD with a C-terminal hexahistidine tag (20) (this tagged AcrD is called “AcrD” in this publication). Cells were grown in Luria-Bertani broth supplemented with ampicillin (100 μg/ml) with rotary shaking at 37°C.

Purification of AcrA and AcrD His-tagged proteins.AcrA, modified by the deletion of the N-terminal lipidation site and by the addition of a hexahistidine tag at the C terminus, was purified as described previously (29). For AcrD purification, an overnight culture of KAM3 containing pTrcHacrD was diluted into 100 vol of LB and incubated until the A₆₀₀ reached 0.4. Expression of AcrD was induced by adding isopropyl-β-d-thiogalactoside to a final concentration of 1 mM. Cells were harvested by centrifugation after 4 h of induction and resuspended in buffer A (20 mM Tris-HCl [pH 7.0], 500 mM NaCl, 5 mM imidazole). The cell suspension was lysed by two passes through a French pressure cell (American Instrument Co.) operating at 10,000 lb/in² in the presence of DNase (100 μg/ml) and EDTA-free protease inhibitor cocktail (Roche). Once unbroken cells were removed by centrifugation, the supernatant was centrifuged at 100,000 × g for 1 h at 4°C. The pellet was resuspended in buffer A containing 1.5% dodecyl maltoside (DM) to solubilize membrane proteins. After the centrifugation step was repeated, the supernatant was used immediately for purification of AcrD, using Talon His-Bind cobalt chelating resin (Clontech) in a hybrid batch-gravity flow procedure as described by the manufacturer with some modifications. The binding and initial washing steps were carried out in buffer A. The last wash of the resin suspension was performed with buffer B (25 mM Tris-HCl [pH 7.0], 200 mM KCl, 10% glycerol, 5 mM imidazole, 0.2% DM). Elution of AcrD was obtained using buffer B containing 100 mM imidazole. Fractions containing AcrD were pooled and concentrated by centrifugation in a Centricon-30 (M_r 30,000-cutoff; Amicon). This solution was used in the reconstitution (see below) without removal of imidazole or DM because their carryover was minimal as the protein concentration was usually high (1.5 to 2 mg/ml).

Production of polyclonal anti-AcrD serum.Purified AcrD was used to generate polyclonal antibodies in rabbit. The antibodies were purified from serum by ammonium sulfate precipitation (70% saturation), and antibodies against non-AcrD proteins were removed by affinity chromatography using the crude extract from JZM320 (acrD::tet) (22) coupled to CNBr-activated Sepharose as described previously (29).

Protein assay and analysis.Protein concentration was determined with the BCA reagent (Pierce) with bovine serum albumin as the standard. Proteins were resolved in sodium dodecyl sulfate (SDS)-7.5% polyacrylamide gel electrophoresis. In some cases, the resolved proteins were transferred electrophoretically to a nitrocellulose membrane (Bio-Rad) in Tris base (20 mM), glycine (150 mM), and methanol (20%) for Western blot analysis. Binding of primary polyclonal antibodies (anti-AcrA [29] or anti-AcrD) was detected with an alkaline phosphate-conjugated anti-rabbit secondary antibody (Sigma). Protein visualization was performed with nitroblue tetrazolium and 5-bromo-4-chloro-3-indolylphosphate (4).

AcrD reconstitution into proteoliposomes.Proteoliposomes were prepared with E. coli lipids (Avanti Polar Lipids) by the octyl-β-d-glucopyranoside (OG) dilution method, described previously (4, 29). Briefly, AcrD (50 μg, usually 25 to 30 μl) was added to 4.5 mg of lipids dispersed in the reconstitution buffer (25 mM HEPES-KOH [pH 7.0], 200 mM KCl, 1 mM dithiothreitol), and the total volume was adjusted to 0.5 ml with the reconstitution buffer and 10% OG so that the final concentration of OG was 1.1%. This solution was diluted with 5 ml of chilled reconstitution buffer, and the proteoliposomes were recovered by centrifugation (4) and were usually resuspended in 100 μl of the reconstitution buffer. As a control, vesicles without AcrD were prepared in the same way. In most cases, pyranine (1 mM) was added to the mixture before dilution so that it would become entrapped in the intravesicular space. In some cases, AcrA (30 μg), aminoglycosides (70 μM), and/or MgCl₂ (1.5 mM) was similarly added to the 0.5-ml mixture before dilution.

Transmembrane H⁺ flux assay.Determination of transmembrane H⁺ flux was performed as described by Zgurskaya and Nikaido (29). Thus, pyranine, a hydrophilic, fluorescent, pH indicator dye, was incorporated into the intravesicular space. These proteoliposomes (up to 20 μl) were added to 2 ml of assay buffer (25 mM HEPES-NaOH [pH 7.0], 200 mM NaCl); valinomycin (10 μM) was added to produce ΔΨ, which becomes converted rapidly to ΔpH in medium containing a high concentration of chloride ion. Fluorescence was followed with a Shimadzu RF-5301 spectrofluorimeter to measure changes in intravesicular pH. When aminoglycosides were added to the external medium, they were present in the assay buffer before the addition of proteoliposomes. Excitation and emission wavelengths were, respectively, 455 and 509 nm.

Uptake experiments.Proteoliposomes (30 μl) were preequilibrated with 370 μl of K⁺-free buffer (25 mM HEPES-NaOH [pH 7.0], 200 mM NaCl) containing [³H]gentamicin (specific radioactivity, 200 mCi/g; American Radiolabeled Chemical) at a final concentration of 10 μM. After the addition of valinomycin, 50-μl samples were removed at different times and diluted into an ice-cold solution containing 0.1 M LiCl and 0.4 mg of poly-l-lysine (Sigma)/ml, and the mixture was filtered on a 0.22 μm-pore-size GSWP Millipore filter. The filter was washed with ice-cold 0.1 M LiCl, dried, and used for determination of radioactivity in a Beckman LS6500 liquid scintillation counter.

Aminoglycoside susceptibility assays.The susceptibility of E. coli to aminoglycosides (obtained from Sigma) was determined in Mueller-Hinton medium (Difco) by the broth microdilution method. The inoculum was 10⁴ cells per ml, and the results were read after overnight incubation at 37°C.

Purification of AcrD and AcrA.The purified AcrD contained a hexahistidine tag at its C terminus (20) and was 95% pure as judged by Coomassie blue staining of SDS-polyacrylamide gels (Fig. 1A, lane 3). This hexahistidine-tagged AcrD was earlier shown to be fully functional in intact E. coli cells (20). Mass spectrometry analysis of its trypsin digest matched the expected pattern for the AcrD protein (data not shown). The purified AcrD preparation was used to produce polyclonal rabbit antibodies.

• Open in new tab
• Download powerpoint

FIG. 1.

Overexpression and purification of AcrD (A) and AcrA (B). Protein samples were loaded onto an SDS-7.5% polyacrylamide gel and stained with Coomassie brilliant blue. Lanes 1, crude extract; lanes 2, column flowthrough fractions; lanes 3, AcrD-6His after cobalt resin batch purification (A) and AcrA-6His after purification on an Ni²⁺ column (B). The rightmost and leftmost lanes in panels A and B, respectively, show molecular mass markers, with molecular masses shown in kilodaltons. (C) Western blot analysis of AcrD reconstituted into proteoliposomes. Here AcrD proteoliposomes were loaded onto an SDS-7.5% polyacrylamide gel and analyzed by immunoblotting with polyclonal anti-AcrD antibodies (left lane). The right lane shows molecular mass markers as for panels A and B.

The purified AcrA (Fig. 1B, lane 3) used in this work corresponds to the modified protein containing a hexahistidine tag at its C terminus and devoid of lipid modification at its N terminus (30). This lipid-deficient version of AcrA (starting at D26 of the native proprotein) was shown to be functionally active (30) and was used in order to avoid possible interactions with the lipid vesicles during AcrD reconstitution assays.

Reconstituted AcrD-containing proteoliposomes were inactive in aminoglycoside transport in the absence of AcrA.An uptake assay performed with intact cells of E. coli showed that aminoglycosides were substrates of AcrD (22). Since AcrD is a member of the RND family of transporters energized by proton motive force, movement of the substrates by the transporter should produce coupled transmembrane movement of H⁺, as was experimentally demonstrated for AcrB (29). We followed the changes in intravesicular pH with the water-soluble fluorescent pH probe pyranine (29), which does not diffuse across the vesicle membrane (3, 12).

When protein-free vesicles containing K⁺ and pyranine were diluted in the assay buffer containing Na⁺ but no K⁺, the addition of valinomycin produced a rapid efflux of K⁺ and generated membrane potential, which was replaced by interior-acid ΔpH in the presence of Cl⁻, as seen by the rapid drop of pyranine fluorescence. Pyranine fluorescence (therefore intravesicular pH) then increased slowly, reflecting the spontaneous leakage of H⁺ (data not shown). In preliminary experiments we tested whether proton leakage was nonspecifically enhanced by the presence of known substrates of AcrD. Addition of aminoglycosides (gentamicin, tobramycin, amikacin, and kanamycin) at 70 μM to the external solution did not alter the rate of proton leakage. In contrast, addition of SDS or deoxycholate at the same concentration caused the instantaneous collapse of the pH gradient, presumably through their detergent activity; thus, these substrates of AcrD (20) were not used in this study.

A transmembrane H⁺ flux assay was then conducted by first adding the aminoglycosides to the external medium bathing AcrD-containing proteoliposomes and then by starting the reaction by generating ΔpH (interior acid) by the addition of valinomycin as described above. In intact cells, the periplasm is more acidic than the cytoplasm, and thus the intravesicular and extravesicular spaces in our experimental system correspond to the periplasm and cytoplasm, respectively. AcrD is expected to be an H⁺/drug antiporter like its close homolog, AcrB (29). If aminoglycosides are captured and transported by the AcrD efflux pump from the cytoplasm, the pump action is thus expected to produce a countermovement of H⁺ (efflux of H⁺ from the vesicles) so that the intravesicular pH will return to the initial, more alkaline value faster than through the spontaneous leakage. Even though orientation of reconstituted AcrD in proteoliposomes is probably random, we expected that imposition of an interior-acid ΔpH would cause the AcrD proteins inserted in the correct orientation (with their periplasmic domain facing inside) to function, pumping in substrate in exchange for a proton(s). However, no enhanced efflux of H⁺ was observed in the presence of 70 μM aminoglycosides in the reaction buffer. We show the case for gentamicin in Fig. 2A, curve 2, which shows unaltered slope from the efflux without any substrate (Fig. 2A, curve 1). These results initially suggested to us that AcrD may not pump out aminoglycosides from the extravesicular space (or cytosol). We therefore carried out the H⁺ flux assay using reconstituted AcrD proteoliposomes containing gentamicin (70 μM) entrapped in the intravesicular space (corresponding to the periplasm). However, again, no stimulation of proton flux was observed (data not shown).

• Open in new tab
• Download powerpoint

FIG. 2.

Effects of aminoglycoside addition to the extravesicular space on proton efflux from AcrD-containing proteoliposomes. AcrD proteoliposomes were reconstituted in a buffer containing 0.2 M KCl and then were diluted in the same buffer containing 0.2 M NaCl instead. ΔpH was generated by the addition of 10 μM valinomycin. Decreased fluorescence of the pyranine corresponds to the interior-acid ΔpH, which dissipated slowly in the absence of drugs. Aminoglycosides were added to the extravesicular space at a concentration of 70 μM. (A) Curve 1, AcrD proteoliposomes alone; curve 2, gentamicin was added to AcrD proteoliposomes; curve 3, gentamicin was added to AcrD proteoliposomes containing AcrA plus Mg²⁺. (B) A single batch of AcrD proteoliposomes with entrapped AcrA and Mg²⁺ was used. Curve 1, without added substrates, as a control; curve 2, anti-AcrD antibodies and gentamicin added to the external medium; curve 3, gentamicin alone added to the external medium. (C) A single batch of AcrD proteoliposomes containing AcrA and Mg²⁺ was used. Curve 1, no-substrate control; curve 2, amikacin (70 μM) added externally; curve 3, gentamicin (70 μM) added externally; curve 4, tobramycin (70 μM) added externally. AU, arbitrary units. Insets in this and subsequent figures show schematically the experimental setup. S, substrate. Black squares, AcrA; grey structures, AcrD.

AcrA is necessary for the transport of aminoglycosides by the reconstituted AcrD proteoliposomes.Recently, Elkins and Nikaido (6) showed that the efficient efflux of amphiphilic substrates by AcrD in intact cells requires AcrA. Because of the apparent absence of activity of reconstituted AcrD in the presence of aminoglycosides, we examined whether the transport activity of AcrD for aminoglycosides required AcrA. When AcrA was added to the reaction buffer surrounding the reconstituted AcrD vesicles and gentamicin was also added to this medium, no enhancement of H⁺ efflux was observed (data not shown). This result was expected since under such conditions AcrA, a periplasmic protein, was present in the extravesicular fluid, corresponding to the cytosol based on the direction of ΔpH. However, when AcrA was entrapped in the intravesicular space of the AcrD-containing vesicles together with 1.5 mM MgCl₂ and ΔpH was created, then a drug-induced proton efflux was observed, for example in the presence of 70 μM gentamicin (Fig. 2A, curve 3; compare with curve 1). Control experiments with AcrA entrapped in vesicles without AcrD unfortunately gave erratic results often showing rapid leakage of protons, possibly due to the presence of the hexahistidine tag on AcrA. Results were reproducible, however, when AcrD-containing proteoliposomes were used to entrap AcrA, and in later experiments vesicles of this type, without the addition of aminoglycosides, were always used as the control. (Each H⁺ efflux experiment reported in this study was successfully reproduced at least five times, although small quantitative variations were occasionally seen.)

Because we could not use control vesicles containing AcrA but without AcrD as mentioned above, it was important to confirm that the enhanced H⁺ efflux (as seen in Fig. 2A, curve 3) was caused by the pumping activity of AcrD. For this purpose we performed the H⁺ efflux assay with the AcrA-containing AcrD proteoliposomes in the presence of purified anti-AcrD polyclonal antibodies (added to the extravesicular space). As seen in Fig. 2B, curve 2, no visible drug-induced proton efflux was observed in the presence of gentamicin as a substrate (compare with the no-substrate control of curve 1). In the positive control without the antibody, gentamicin again produced significant acceleration of H⁺ efflux (curve 3). (The rate of H⁺ movement, especially that of H⁺ leakage, was strongly influenced by the particular batch of lipids used for reconstitution and other variables difficult to define. Thus, the slopes should be compared ideally within the same experiments done by using one batch of vesicles, or at least by using the vesicles made side by side on the same day. We cannot compare slopes of curves from different experiments, for example, Fig. 2A, curve 3, with Fig. 2B, curve 3.)

While the addition of 100 μM gentamicin to the external medium resulted sometimes in aggregation of vesicles, concentrations of 80, 70, 50, and 30 μM were able to produce drug-induced proton efflux (data not shown). Similar enhancement of H⁺ efflux was observed with 70 μM amikacin (Fig. 2C, curve 2, compare with curve 1 [no substrate] and with curve 3 [gentamicin control]), tobramycin (Fig. 2C, curve 4, compare with curve 1), or kanamycin (data not shown). The results shown in Fig. 2C indicate that AcrA in the intravesicular space activated AcrD and that the AcrA-AcrD complex acted as an antiporter, catalyzing the efflux of H⁺ coupled presumably to the influx of aminoglycosides into vesicles.

The presumed intravesicular accumulation of aminoglycosides under these conditions was tested by uptake experiments using radiolabeled gentamicin. AcrD proteoliposomes with entrapped AcrA and Mg²⁺ indeed accumulated more gentamicin than the control liposomes (Fig. 3). Using the intravesicular volume of 0.04 μl per 1 μg of AcrD (see Discussion), the accumulation of 3 to 6 pmol of gentamicin accumulated per μg of AcrD translates into an internal concentration 75 to 150 μM, about an order of magnitude higher than the concentration of gentamicin (10 μM) in the assay medium.

• Open in new tab
• Download powerpoint

FIG. 3.

Accumulation of [³H]gentamicin in AcrD-containing proteoliposomes. AcrD proteoliposomes were made in KCl buffer so that AcrA and Mg²⁺ were entrapped within. These were diluted into NaCl buffer containing 10 μM [³H]gentamicin as described in Materials and Methods (▪). The results shown are averages of five independent experiments. Liposomes without AcrD and AcrA served as a control (▴). At time zero 10 μM valinomycin was added to produce a proton motive force.

We also tested if AcrA was required for AcrD-mediated aminoglycoside efflux in intact cells, because an earlier study from our laboratory (6) examined the AcrA requirement only for amphiphilic substrates. Indeed, a ΔacrA mutant of E. coli showed levels of susceptibility to aminoglycosides in Mueller-Hinton medium equivalent to those of the isogenic ΔacrD strain (Table 1). (HNCE1b, lacking AcrA in addition to AcrB and AcrD, appeared to be slightly more susceptible than HNCE1a, lacking AcrB and AcrD alone. This could be because in an AcrBD-negative background the promiscuity of AcrA, which was shown to work together with other pumps such as YhiV [7], becomes detectable.)

The AcrA-AcrD complex can capture substrates from within the vesicles.Because the very hydrophilic aminoglycosides are not expected to cross the lipid bilayer readily, we can entrap these compounds inside the vesicles and test if the AcrA-AcrD complex is capable of capturing them from this space, which corresponds to the periplasm according to the direction of ΔpH. When 70 μM gentamicin or amikacin was entrapped in the intravesicular space of the reconstituted AcrD proteoliposomes in the presence of AcrA and 1.5 mM MgCl₂, the presence of the drug enhanced proton efflux strongly (Fig. 4, curve 2) in comparison with that of the control AcrD proteoliposomes containing only AcrA and Mg²⁺ (curve 1). Entrapment of amikacin showed similar stimulation (data not shown). AcrD proteoliposomes containing drugs and Mg²⁺, but not AcrA, did not show enhanced H⁺ efflux, indicating again that AcrA is needed for the activation of AcrD transporter activity. Thus, aminoglycosides present in the intravesicular space, corresponding to the periplasmic space, enhanced proton movement by the AcrD transporter if AcrA was present. Although we cannot create the barrier equivalent to the outer membrane in our reconstituted system, we believe that in intact cells the proton flux catalyzed by AcrD is energizing the efflux of aminoglycosides from the periplasmic space in intact cells with the help of AcrA and TolC.

• Open in new tab
• Download powerpoint

FIG. 4.

Effect of the presence of aminoglycosides in the intravesicular space on proton efflux from AcrD proteoliposomes. Experiments were carried out as for Fig. 2, using proteoliposome preparations made side by side on the same day, except that the substrate was entrapped within one batch of vesicles. Curve 1, AcrD vesicles with entrapped AcrA and Mg²⁺, as the control; curve 2, AcrD vesicles with entrapped AcrA, Mg²⁺, and gentamicin (70 μM).

Streptomycin as a substrate for the AcrD pump.Surprisingly, when streptomycin was used in our system as a substrate for the reconstituted AcrD, it induced a proton efflux response only when the antibiotic was entrapped in the intravesicular space together with AcrA and Mg²⁺ (Fig. 5, curve 3). Addition of streptomycin to the external medium surrounding the AcrD proteoliposomes containing AcrA and Mg²⁺did not alter the rate of H⁺ efflux in a detectable manner (Fig. 5, compare curve 2 with curve 1). It is likely that AcrD pumps out streptomycin preferentially from the periplasm.

• Open in new tab
• Download powerpoint

FIG. 5.

Effect of streptomycin on proton efflux from AcrD proteoliposomes. Experiments were carried out as for Fig. 2 and 4. Two kinds of proteoliposomes used were made side by side on the same day. Curve 1, AcrD proteoliposomes containing AcrA and Mg²⁺ (but no substrate) as the control; curve 2, streptomycin (70 μM) was added to the extravesicular space of AcrD proteoliposomes containing AcrA and Mg²⁺; curve 3, AcrD proteoliposomes containing AcrA, Mg²⁺, and 70 μM streptomycin. AU, arbitrary units.