Unfolding study of a trimeric membrane protein AcrB

Abstract

The folding of a multi-domain trimeric α-helical membrane protein, Escherichia coli inner membrane protein AcrB, was investigated. AcrB contains both a transmembrane domain and a large periplasmic domain. Protein unfolding in sodium dodecyl sulfate (SDS) and urea was monitored using the intrinsic fluorescence and circular dichroism spectroscopy. The SDS denaturation curve displayed a sigmoidal profile, which could be fitted with a two-state unfolding model. To investigate the unfolding of separate domains, a triple mutant was created, in which all three Trp residues in the transmembrane domain were replaced with Phe. The SDS unfolding profile of the mutant was comparable to that of the wild type AcrB, suggesting that the observed signal change was largely originated from the unfolding of the soluble domain. Strengthening of trimer association through the introduction of an inter-subunit disulfide bond had little effect on the unfolding profile, suggesting that trimer dissociation was not the rate-limiting step in unfolding monitored by fluorescence emission. Under our experimental condition, AcrB unfolding was not reversible. Furthermore, we experimented with the refolding of a monomeric mutant, AcrB_Δloop, from the SDS unfolded state. The CD spectrum of the refolded AcrB_Δloop superimposed well onto the spectra of the original folded protein, while the fluorescence spectrum was not fully recovered. In summary, our results suggested that the unfolding of the trimeric AcrB started with a local structural rearrangement. While the refolding of secondary structure in individual monomers could be achieved, the re-association of the trimer might be the limiting factor to obtain folded wild-type AcrB.

Introduction

After their translation, proteins need to fold into and maintain a three-dimensional structure to function. The folding process and thermodynamic stability of proteins have been an actively pursued area of study due to their importance in virtually all life processes. The thermodynamic stability measurement of integral membrane protein remains to be a challenging task and study on the mechanism of their folding is scant.^1–4 To quantitatively determine the thermodynamic stability of a protein, its reversible unfolding and refolding has to be achieved.^4–6 The first α-helical membrane protein that has been successfully refolded from an sodium dodecyl sulfate (SDS) denatured state is bacteriorhodopsin.^7,8 Since then, refolding of a few additional α-helical membrane proteins has been reported, including diacylglycerol kinase,⁹ a potassium channel KcsA,¹⁰ and the lactose permease LacY.¹¹

Refolding is challenging for helical membrane proteins. The approach taken to unfold a protein is crucial for its refolding efficiency. Methods suitable for soluble proteins such as thermal and acid/base denaturation usually lead to irreversible precipitation or aggregation of α-helical membrane proteins due to their large hydrophobic core.^4,12 So far, chemical unfolding using urea, SDS, and trifluoroethanol (TFE) is the only viable approach demonstrated to enable successful refolding of α-helical membrane proteins.^13–15 Urea exerts its effect directly, by binding to the protein, or indirectly, by altering the solvent environment, both resulting in the solvation of the hydrophobic core of the protein,¹⁶ while the mechanism of denaturation caused by SDS remains a topic of debate.^6,17,18 The far ultraviolet (UV) circular dichroism (CD) analysis suggests that membrane proteins denatured by SDS tend to maintain a high level of helical structure.^19,20 This state is arguably a better representative of the unfolded state of helical membrane proteins.^5,21 An experimental readout that is sensitive to the conformational change of protein during unfolding is another critical factor of the stability measurement. Methods that have been intensively used include the far UV CD spectrum which is sensitive to the secondary structure variation, and the intrinsic fluorescence spectrum that reveals the overall conformational change of a protein.^19,20,22

In this study, we investigated the unfolding and possibility of refolding of an Escherichia coli inner membrane protein AcrB. Although many attempts have been made to study the folding of α-helical membrane proteins, these folding studies focus on membrane proteins with relatively simple structures, such as the small multidrug resistance transporter EmrE,²³ the human peripheral myelin protein 22,²⁴ the ATP-binding cassette transporter BtuCD,²⁵ and the sugar transport protein GalP²⁶ and LacY,^11,27 all are helix bundles containing a handful of transmembrane helices. Report on the folding study of multimeric and multidomain helical membrane proteins remains scarce. Escherichia coli AcrB is an inner membrane transporter that forms a functional complex with a membrane fusion protein AcrA and an outer membrane channel TolC in the cell membrane to actively transport various compounds out of cell.^28,29 The crystal structure of AcrB^30,31 has been obtained a decade ago. It is an obligate trimer that contains both a transmembrane domain and a large periplasmic domain (Fig. 1). Each monomer contains 12 transmembrane helices and a periplasmic domain with mixed secondary structure, formed by two long loops between transmembrane helices 1 and 2, and helices 7 and 8. Thermal unfolding monitored using CD spectroscopy at 222 nm has been measured in a previous study.³² The protein retained a high helical content even when the temperature rose up to 98°C. In addition, a clear transition point was lacking. In this study, we examined AcrB unfolding induced by two chemical denaturants, urea and SDS. Intrinsic fluorescence emission and CD signals of the protein were monitored to follow the unfolding and refolding processes. While wild-type AcrB failed to refold reversibly under the experimental conditions tested in this study, the overall secondary structure composition of its monomeric mutant, AcrB_Δloop, could be regained after refolding. The unfolding curve could be fitted with a two-state unfolding model, indicating the lack of a folding intermediate state.

Figure 1.

Structure of AcrB (created from 2GIF.pdb⁴⁰ using Pymol⁴¹). Top view (left) and side view (right, with the subunit at the back removed for clarity) of an AcrB trimer. Positions of domains as discussed in the text are labeled in the side view. The two lines marked the position of the cell membrane.

Results

Chemical denaturation of trimeric AcrB monitored by intrinsic fluorescence

A reversible unfolding/refolding process is a prerequisite for obtaining thermodynamic parameters of protein stability.⁴ To study the stability of AcrB, we examined its unfolding in urea and SDS, and monitored the decrease of intrinsic fluorescence intensity during unfolding. Protein unfolding usually leads to the transition of fluorescent residues to shift from the protected hydrophobic core to an exposed hydrophilic environment, resulting in a decrease of fluorescence signal and sometimes a red shift of the emission peak.

When AcrB unfolded in urea, a dramatic decrease of fluorescence intensity was observed [Fig. 2(A)]. Therefore, we chose to monitor the decrease of the maximum fluorescence emission as an indicator of protein unfolding. Titration curve in urea [Fig. 2(B)] displayed a sigmoid profile with a transition point at around 3.0M and the protein completely unfolded when urea concentration reached 5.0M.

Figure 2.

Chemical denaturation of AcrB monitored using the intrinsic fluorescence. A: Spectra of AcrB in the absence (solid line) and the presence of (dashed line) of 6M urea. B: Unfolding of AcrB induced by urea titration. C: Spectra of AcrB in the absence (solid line) and the presence of SDS (dashed line, SDS/DDM molar ratio 2.4). D: Unfolding of AcrB induced by the titration of SDS (black squares). l-Tryptophan (open squares) was used as a control to confirm the lack of non-specific effect of SDS on the fluorescence emission. Each experiment was repeated at least three times, and the average values and standard deviations are shown.

A significant decrease of the fluorescence signal was also observed when we performed a similar titration using SDS [Fig. 2(C)]. The transition point occurred at the SDS/n-dodecyl β-d-maltoside (DDM) molar ratio of around 1.0 [Fig. 2(D)]. A small fluorescent molecule l-tryptophan was titrated similarly with SDS to serve as a control of potential non-specific fluorescence change. As revealed in Figure 2(D), the presence of SDS had little effect on the fluorescence of l-tryptophan, indicating that the change of buffer condition upon the addition of SDS did not have a non-specific effect on the fluorescence emission. Therefore, the decrease in AcrB fluorescence emission during the titration process was caused by the protein structure change. Titration results revealed that the protein fully unfolded when the SDS/DDM molar ratio reached 2.4.

Contribution of signal change from trimer dissociation

Dissociation of AcrB trimer in the presence of SDS was confirmed using blue native polyacrylamide gel electrophoresis (BN-PAGE). Only one trimer band could be observed when a fresh AcrB sample was analyzed using BN-PAGE [Fig. 3(A), Lanes 1]. In the presence of SDS at the molar ratio of 2.4, the trimer band was completely converted into a monomer band, indicating that all AcrB trimers dissociated into monomers [Fig. 3(A), Lane 2]. To confirm the observed change of migration was not due to the nonspecific association of SDS, we conducted a control experiment using a monomeric AcrB mutant created in a previous study, AcrB_Δloop.³³ In AcrB_Δloop, the long loop that is critical for the AcrB trimerization is truncated, which leads to trimer dissociation. However, the overall secondary and tertiary structure of the mutant was comparable to that of the wild type AcrB.³³ As shown in Figure 3(A), Lanes 3 and 4, the addition of SDS did not shift the migration of AcrB_Δloop in the gel, indicating the observed change in Lane 2 was due to trimer dissociation.

Figure 3.

AcrB trimer dissociation and SDS-induced unfolding. A: Representative BN-PAGE gel image of wild type AcrB samples with (Lane 2) or without (Lane 1) incubation with SDS (SDS/DDM molar ratio 2.4). AcrB_Δloop samples were treated similarly and used as controls (Lanes 3 and 4). Positions of monomer (M) and trimer (T) as well as molecular weight markers are labeled. B: SDS titration of wild type AcrB (black squares) and AcrB_V225C/A777C (gray triangles). C: SDS titration of wild type AcrB before (black squares) or after (open squares) the reduction of DTT. D: SDS titration of AcrB_V225C/A777C before (gray triangles) or after (open triangles) the reduction of DTT. Each experiment was repeated at least three times and the average values and standard deviations are shown.

To examine the contribution of trimer dissociation and monomer unfolding to the observed change of fluorescence, we monitored the unfolding of a mutant, AcrB_V225C/A777C. AcrB_V225C/A777C contains an inter-subunit disulfide bond between each pair of neighboring subunits, and exists as a covalently linked trimer.³⁴ Formation of the disulfide bond between V225C and A777C of adjacent monomers was complete as revealed by the lack of monomer band when the protein was analyzed on non-reducing SDS-PAGE gel.³⁵ These inter-cellular disulfide bonds have no effect on the AcrB activity, indicating a lack of negative impact on protein structure.³⁴ We compared the titration curve of AcrB_V225C/A777C with the curve of wild-type AcrB, and their titration curves under a reducing condition with a non-reducing condition [Fig. 3(B–D)]. Incubation in the presence of 0.5 mM of dithiothreitol (DTT) at room temperature for 2 h completely reduced disulfide bonds in AcrB_V225C/A777C trimer (data not shown). According to Figure 3(C), the overlapped titration curves of wild-type AcrB in the absence and the presence of DDT confirmed that DTT had no influence on protein structure except for reducing disulfide bonds. The loss of disulfide bonds in AcrB_V225C/A777C trimer had little effect on protein stability [Fig. 3(D)]. Therefore, AcrB monomer unfolding was the primary cause of the observed sharp decrease in the intrinsic fluorescence intensity.

Unfolding of a triple AcrB mutant

As discussed above, the intrinsic fluorescence of a protein can reflect its conformational state. Trp residues are the major contributor of protein intrinsic fluorescence. There are nine Trp residues in the sequence of AcrB, three in the transmembrane domain (W13, W515, and W895) and six in the periplasmic domain (W187, W634, W754, W789, W809, and W859). We chose to mutate all three Trp in the transmembrane domain into Phe to investigate the effect on the observed unfolding profile. To confirm that the mutations did not drastically affect AcrB structure, we determined the activity of the mutant by measuring the minimal inhibitory concentration (MIC) of an AcrB knockout strain BW25113ΔacrB harboring a plasmid encoding AcrB_{W13F/W515F/W895F} (pQE70-AcrB_{W13F/W515F/W895F}). The same strain harboring the empty vector (pQE70) or plasmid encoded wild-type AcrB (pQE70-AcrB) was used as the negative and positive control, respectively. Two well-established AcrB substrates were tested, erythromycin and novobiocin. For erythromycin, the MIC of the strains containing wild type, mutant, or no AcrB was 160 µg/mL, 160 µg/mL, and 5 µg/mL, respectively. For novobiocin, the MIC of the three strains was 80 µg/mL, 40 µg/mL, and 10 µg/mL, respectively. For both substrates tested, the mutant retained complete or partial activity, indicating that the overall structure of the protein was largely intact. The replacement of three Trp with Phe residues led to a decrease of the fluorescence intensity by ∼30%, and the peak of the spectrum shifted to a slightly longer wavelength [Fig. 4(A)]. Next, SDS induced unfolding was measured [Fig. 4(B)]. The unfolding profile of the triple mutant was very similar to the profile of wild type AcrB. Removal of three Trp from the transmembrane domain did not lead to an observable shift of the unfolding profile, indicating that the observed signal change was actually largely contributed by Trp residues from the periplasmic domain.

Figure 4.

SDS-induced AcrB_{W13F/W515F/W895F} unfolding. A: A comparison of fluorescence spectra of wild type AcrB (solid line) and mutant AcrB_{W13F/W515F/W895F} (dashed line). The spectra were normalized to reveal the peak shift. Original concentration-adjusted spectra are shown as an inset. B: Unfolding of AcrB_{W13F/W515F/W895F} induced by the titration of SDS (open squares). Unfolding of wild type AcrB is also shown as a control (gray squares). Each experiment was repeated at least three times, and the average values and standard deviations are shown.

Chemical denaturation of trimeric AcrB monitored by CD spectroscopy

To further examine the structural change of protein in the presence of SDS, far UV CD spectra were collected for AcrB sample in the absence and presence of SDS at a SDS/DDM molar ratio of 2.4. Changes in both peak height and wavelength were observed in the CD spectra, revealing a shift of the secondary structure composition and arrangement [Fig. 5(A)]. Effort to deconvolute the CD spectra into individual secondary components using online software provided by the DICHROWEB³⁶ server was not successful, likely due to the lack of an appropriate reference set for membrane proteins. The disruption of secondary structure components during SDS titration was monitored using CD. The ratio of mean residue ellipticity at 222 nm and 208 nm was used to reflect the secondary structure change [Fig. 5(B)]. Similar to that monitored using intrinsic fluorescence, structural change was complete at a SDS/DDM molar ratio of 2.4.

Figure 5.

SDS-induced denaturation monitored using CD. A: Spectra of purified AcrB in the absence (solid line) and the presence (dashed line, SDS/DDM molar ratio 2.4) of SDS. B: AcrB titration with SDS monitored by the ratio of ellipticities at 208 nm and 222 nm. C: Comparison of unfolding profiles. Comparison of the process of unfolding monitored using CD (gray squares) and fluorescence intensity (black square). The curves showed the fitting of the CD and fluorescence intensity data using a simple two-states unfolding model. Each experiment was repeated at least three times and the average value and standard deviation are shown.

Comparison of AcrB unfolding plots monitored using fluorescence and CD spectroscopy

While both unfolding plots revealed that the structure change (unfolding) was complete at a SDS/DDM molar ratio of 2.4, the transition points were significantly different [Fig. 5(C)]. Furthermore, both titration plots could be fitted with a sigmoid curve describing a simple two-state unfolding model from a folded state directly to an unfolded state [Fig. 5(C)]. The transition point of wild-type AcrB unfolding was calculated to be at SDS/DDM molar ratios of 0.94 ± 0.03 and 1.32 ± 0.09 for CD and fluorescence emission, respectively. Comparison between the plots suggested that the rearrangement of protein secondary structure occurred prior to the global unfolding of the tertiary structure.

Refolding of chemically denatured AcrB

To quantitatively determine the thermodynamic parameters of AcrB, reversible refolding needs to be achieved. We first unfolded AcrB in the presence of 6M urea, and then diluted the sample to reduce the concentration of urea. As shown in Figure 6(A,B), the fluorescence intensity begun to increase when the concentration of urea was reduced to lower than 2M, and 57% fluorescence were recovered when the urea concentration was reduced to 1M. The presence of two separated unfolding and refolding paths and the low recovery efficiency revealed that AcrB did not refold reversibly from the urea-denatured state under our experimental condition. We have also collected the CD spectra of AcrB unfolded in urea (dotted line), and then after dialyzed overnight (gray line) [Fig. 6(C)]. The secondary structure composition of the protein was not regained.

Figure 6.

Refolding of AcrB. A: Unfolding (black) and refolding (gray) profiles of AcrB in urea. B: Fluorescence emission spectra of AcrB in the native state (black line), unfolded (dashed line, in 6.0M urea), and after refolding (gray line). C: CD spectra of AcrB in the native state (black line), unfolded state (dashed line, in 6.0M urea), and after refolding (gray line) in urea. D: Unfolding (black) and refolding (gray) profiles of AcrB in SDS. E: Fluorescence emission spectra of AcrB in the native state (black line), unfolded (dashed line, SDS/DDM molar ratio 2.4), and after refolding (gray line). F: CD spectra of AcrB in the native state (black line), unfolded state (dashed line, SDS/DDM molar ratio 2.4), and after refolding (gray line) in SDS.

To refold the SDS-denatured AcrB, we reduced the molar ratio of SDS in SDS/DDM mixed micelles by adding aliquots of concentrated DDM solution. As shown in Figure 6(D,E), the refolding curve did not overlap with the unfolding curve, with around 25% fluorescence recovery at the lowest SDS/DDM molar ratio tested (0.07). The CD spectrum of refolded sample were collected and compared with the spectra of native and unfolded AcrB samples [Fig. 6(F)]. Again, the secondary composition of the protein was not fully recovered in the refolding sample. It was clear that the structure of the protein was different from the native-like conformation. We have also attempted to refold the protein under different conditions, including varying the pH, salt concentration, lipid vesicle concentration, and detergent composition. None of these conditions led to the successful refolding of the protein to the original conformation (data not shown).

Next, we experimented with the refolding of an AcrB mutant, AcrB_Δloop. As mentioned earlier, AcrB_Δloop exists as a monomer, but has similar secondary and tertiary structure as a monomeric subunit in a wild-type trimer.³³ Since the complete refolding of wild-type AcrB involves trimer association, we speculate that the refolding of a monomeric AcrB mutant might be easier to achieve. Unlike the SDS induced unfolding of the wild-type protein, the unfolding of AcrB_Δloop was not as cooperative [Fig. 7(A)]. However, the unfolding was partially reversible when the SDS/DDM molar ratio was reduced upon the addition of DDM. The CD spectrum of the refolded sample superimposed well with that of the original folded protein, while the fluorescence spectrum was not fully restored [Fig. 7(B,C)].

Figure 7.

Refolding of AcrB_Δloop. A: SDS induced unfolding (black diamonds) and refolding (gray diamond) profiles of AcrB_Δloop. B: Fluorescence emission spectra of AcrB_Δloop in the native state (black line), unfolded (dashed line), and after refolding (gray line). C: CD spectra of AcrB_Δloop in the native state (black line), unfolded state (dashed line), and after refolding (gray line).

Discussion

In this study, we monitored the chemical denaturation of trimeric AcrB in the presence of urea and SDS. Both trimer dissociation and monomer unfolding could potentially lead to a change of the intrinsic fluorescence emission. However, we found that the presence of inter-subunit disulfide bonds, which strengthened AcrB trimer stability by covalently linking the three subunits, had little effect on the transition point of AcrB unfolding monitored using fluorescence intensity. There are two potential reasons that may lead to this observation. First, trimer dissociation could be spectroscopically silent. Second, trimer dissociation could have occurred together with the global unfolding of individual subunits. To further investigate the unfolding process, we monitored the folding process using CD. The unfolding profiles monitored using CD was different from that monitored using fluorescent intensity. Even though SDS cannot effectively unfold α-helices in the transmembrane domain, it causes a peak shift of the CD spectrum and has been used as a probe of secondary structure change during α-helical membrane protein unfolding.^14,15,22 It is notable in our results that the initiation and transition points of unfolding monitored using CD occurred at a significantly lower SDS concentration than that monitored using fluorescence emission spectroscopy. Therefore, the unfolding of AcrB trimer appeared to start with the structural rearrangement of monomers, and possibly the concurrent dissociation of trimers. The fluorescent intensity change indicates the disruption of the hydrophobic core and exposure of aromatic residues. Based on our result, the change of fluorescence signal was largely contributed by the unfolding of the soluble domain, which could be fitted nicely with a two-state model. The unfolding of the transmembrane domain may not generate a detectable fluorescence signal change, since the change of microenvironment for Trp residues in the transmembrane domain upon unfolding is not likely as dramatic as Trp residues in the soluble domain.

We also attempted to refold the chemically denatured AcrB to quantitatively determine the thermodynamic stability of folding for trimer AcrB. However, under our experimental condition, neither urea nor SDS denatured AcrB can be reversibly refolded.

Interestingly, an AcrB monomeric mutant, AcrB_Δloop, could regain its secondary structure composition after SDS induced unfolding. The protein was not fully refolded into the native conformation as judged by the fluorescence spectrum. The fine-tuning of the tertiary structure during AcrB in vitro refolding might require trimerization. We speculate that trimerization, which is the last hurdle for the wild-type AcrB to refold into a native-like structure, might be difficult to achieve in an artificial refolding condition. The complete refolding of AcrB might require a condition more closely mimicking the native cell membrane, in which multiple subunits are accessible to each other and the transmembrane domain is stabilized in a lipid bilayer.

The unfolding process of AcrB trimer was clearly different from the observed unfolding process of the tetrameric GlpF, which occurs via a stable dimeric intermediate state.³⁷ A larger oligomeric membrane protein, the heterotetrameric BtuCD, has been refolded in detergent micelles.²⁵ The protein complex consists of two transmembrane protein subunits, BtuC, and two cytoplasmically located nucleotide-binding protein subunits, BtuD. Results from the refolding study support the idea of cooperative folding and assembly of the constituent protein subunits of the BtuCD transporter, which is consistent with the observed correlation between subunit unfolding and dissociation during the unfolding of an AcrB trimer. Different from AcrB, the structural organization in BtuCD is modular. Therefore, there is a certain degree of flexibility to allow independent folding of BtuC in detergent micelles and BtuD in aqueous solution. In contrast, the sequence segments that fold into the transmembrane domain and segments that fold into the periplasmic domain in AcrB are interdigitated. Folding of these two domains is likely more tightly correlated and reliant on each other. It has been widely accepted that the unfolded state of a transmembrane helix bundle remains to be highly helical. In the case of AcrB, we speculate that the unfolded state of the protein is consisted of a loosely packed transmembrane helix bundle tied to the more disordered periplasmic loops. The anchoring effect of the transmembrane component could actually be a facilitating factor during the refolding of the periplasmic domain, which subsequently brings the transmembrane helices back to their native-like positions.

Materials and Methods

Plasmid construction, protein expression, purification, and activity assay

Plasmid pQE70-AcrB, pQE70-AcrB_Δloop, and pQE70-AcrB_V225C/A777C were constructed in earlier studies.^32,33 Plasmid encoding the triple mutant AcrB_{W13F/W515F/W895F} was constructed using the QuikChange Site Directed Mutagenesis Kit (Agilent Technologies) with pQE70-AcrB as the template. Protein was expressed and purified as described.³⁸ In the elution step, the elution buffer contained 0.5M imidazole, 0.03% DDM, 20 mM Na-phosphate, 0.3M NaCl (pH 7.9). Protein samples were dialyzed overnight in the same buffer in the absence of imidazole at 4°C overnight. The same buffer was used throughout this study unless otherwise noted. Protein concentration was determined using the Pierce BCA Protein Assay Kit (Thermal Scientific, TX). Protein activity was measured using a drug susceptibility assay as described.³⁸

Unfolding of AcrB

Protein unfolding was initiated by titrating concentrated urea (8.0M) or SDS (0.2% or 2%, wt/vol) into freshly purified AcrB in the dialysis buffer. Fluorescence emission spectrum ranging from 300 nm to 400 nm was recorded at each titration point on a LS-55 fluorescence spectrometer (PerkinElmer, Inc., Waltham, MA) with the excitation wavelength of 280 nm. The maximum fluorescence intensity of each spectrum was normalized using Eq. (1) and plotted as a function of urea molar concentration or DDM/SDS molar ratio. Titration plot of l-tryptophan with SDS solution as a control was obtained similarly.

(1)

in which, y denotes the fluorescence intensity, and and are the maximum and minimum values, respectively.

To reduce the disulfide bonds in AcrB_V225C/A777C, purified protein was divided into two aliquots: one incubated in the dialysis buffer containing 0.5 mM DTT at 25°C for 2 h to reduce disulfide bonds, the other incubated in the absence of DDT under the same condition. Wild-type AcrB was treated similarly and used as the control. Plots of normalized fluorescence versus DDM/SDS molar ratio were obtained as described above.

BN-PAGE analysis and far UV CD spectroscopy

Purified protein was incubated in a buffer solution in the absence or presence of SDS with molar ratio of 2.4 at 25°C for 20 min and then subjected to BN-PAGE analysis as described.^35,39

Purified AcrB in buffer solution was titrated with 0.2% (wt/vol) SDS solution at 25°C. CD spectra in the range of 190 nm and 250 nm were collected using a JASCO J-810 spectrometer. The ratio of the mean residue ellipticity at 222 nm and 208 nm was normalized using Eq. (1) and plotted as a function of DDM/SDS molar ratio.

Refolding of AcrB

To refold protein denatured by urea, purified AcrB was incubated in the dialysis buffer containing 6M urea at 25°C for 20 min to be completely unfolded, followed by titration with the urea-free buffer. To refold SDS denatured protein, the buffer solution containing 10% DDM was titrated into the unfolded protein sample containing SDS at the molar ratio of 11.8. Fluorescence emission spectra and CD spectra were recorded and analyzed to obtain the refolding curve as described above.

Glossary

BN-PAGE

blue native polyacrylamide gel electrophoresis

CD

circular dichroism

DDM

n-Dodecyl β-d-maltoside

DTT

dithiothreitol

MIC

minimal inhibitory concentration

SDS

sodium dodecyl sulfate

UV

ultraviolet