Vanderbilt Institute of Chemical Biology



Discovery at the VICB







Dynamics of a Multidrug Resistance Transporter


By: Carol A. Rouzer, VICB Communications
Published:  January 26, 2016



Double electron-electron resonance studies of the EmrE small multidrug resistance transporter from E. coli provides key insights into its transport mechanism.


An important mechanism by which bacteria evade the effects of antibiotics is through expression of membrane proteins that export potentially toxic substances out of the cell. The small multidrug resistance (SMR) transporters are a class of proteins that carry out this function by harnessing the energy of the proton gradient across a bacterial cell’s inner membrane to expel hydrophobic cations. A prototypical SMR transporter is EmrE from Escherichia coli. Like other SMR transporters, EmrE is a homodimer of subunits, each comprising four hydrophobic transmembrane helices (TM1 - TM4). TM1, TM2, and TM3 form a substrate binding pocket deep within the membrane while TM4 plays a role in dimer formation. Extensive prior studies have revealed the key role of glutamate-14, which is located in EmrE’s substrate binding pocket and alternately binds substrate and protons. Although this finding suggests a mechanism for proton-dependent substrate transport through mutually exclusive binding to glutamate-14 (Figure 1a), the structural basis for the mechanism remains poorly defined. To address this question, Vanderbilt Institute of Chemical Biology members Hassane Mchaourab and Jens Meiler used double electron-electron resonance spectroscopy (DEER) in combination with molecular modeling to probe the conformational changes that occur in EmrE during proton and substrate binding [R. Dastvan et al., (2016) Proc. Natl. Acad. Sci. U.S.A., published online January 19, DOI:10.1073/pnas.1520431113].



FIGURE 1. (a) Simplified mechanism of transport by EmrE. A proton enters from the relatively acidic periplasmic space, causing a conformational change that leads to release of the proton into the cytoplasm. Drug enters from the cytoplasm, resulting in another conformational change that leads to release of drug into the periplasm. (b) Crystal structure of EmrE with TPP bound. The two antiparallel subunits exhibit different conformations. Transport is proposed to occur through interconversion of the subunits between the two conformations. Image reproduced by permission from Macmillan Publishers Ltd, from E. A. Morrison, et al., (2012) Nature, 481, 21545. Copyright 2011.


In the only available low resolution crystal structure of EmrE, the protein is complexed with a cationic substrate, tetraphenylphosphonium (TPP). The structure shows the two subunits in an antiparallel orientation with different conformations. This structure has led to the hypothesis that transport of protons and substrates by EmrE occurs through interconversion of the subunits between the two conformations (Figure 1b). Such an interconversion would expose the opening into the active site alternately between the cytoplasmic and periplasmic side of the membrane. However, prior work in the Mchaourab lab [S. T. Amadi, et al., (2010) J. Biol. Chem., 285, 26710] called this hypothesis into question. Their structural dynamics studies of EmrE in liposomes revealed discrepancies with the X-ray crystallography data, suggesting that the structure of the protein in crystals may not reflect its structure in membranes. Furthermore, the investigators realized that a complete cycle of the pump required proton-bound conformations in addition to substrate-bound ones. This led them to conduct a more detailed investigation of EmrE structure and dynamics using DEER.


A complete DEER analysis requires that researchers first construct a mutant protein in which all cysteine residues have been converted to alanine or serine. They then must design and express a series of mutant proteins, each containing one or more cysteine residues located at carefully selected sites that are predicted to move with respect to each other during a conformational transition. Once this is accomplished, the mutant proteins are labeled with a sulfhydryl-targeted spin-label. The Mchaourab lab used MTSSL (1-oxyl-2,2,5,5-tetramethyl-Δ3-pyrroline-3-methyl) methanethiosulfonate) for this purpose. After careful verification that the structural changes did not significantly affect protein function, the spin-labeled proteins are ready for DEER measurements. These enable the researcher to monitor changes in the distance between two spin-labeled sites in the protein. For investigation of EmrE dynamics, the Mchaourab lab already had a library of DNA constructs of single cysteine mutants available from their prior work. Each protein contained a single label in one of the TMs or a loop connecting the TMs. Because the subunits adopt an antiparallel orientation, the label on each subunit is localized on opposite ends of the dimer, providing insight into the movement of each segment of the protein between the two monomers (Figure 2).




FIGURE 2. Location of representative spin-labels in the TM regions used in the study. For each spin label pair, DEER data measuring the probability distribution (P(r)) for the distance (r) between the labels is shown. Distinct distributions between intermediates suggest the existence of multiple conformations. Broad distributions, especially under pH 8 conditions, suggest protein flexibility. Sharper peaks, indicating a more restricted set of conformations, are seen upon the switch to pH 5 or, particularly, addition of TPP. Image reproduced by permission from R. Dastvan et al., (2016) Proc. Natl. Acad. Sci. U.S.A., published online January 19, DOI:10.1073/pnas.1520431113. Copyright 2016, R. Dastvan, et al.


To attempt to capture the full range of protein conformations, the investigators generated DEER data under three sets of conditions. They used pH 5 to mimic the acidic environment of the bacterial periplasmic space, conditions that were expected to result in protonation of EmrE. Conversely, they chose pH 8 to mimic cytoplasmic conditions that were expected to generate an unbound, apo form of the protein. Addition of TPP at pH 8 yielded a substrate-bound conformation.


Because proteins are not rigid, static structures, DEER data are actually distributions that provide the probability that the two labels will be at a particular distance from each other in a given state. A broad distribution suggests conformational flexibility between the two labeled sites, while a narrow distribution suggests rigidity. As expected, the DEER data confirmed that the protein adopted distinct conformations under each of the three experimental conditions. Data for the TPP-bound conformation agreed with their prior work, indicating discrepancies with the crystal structure. Furthermore, the data ruled out the hypothesis that a simple alternation of the subunits between two different conformations could explain EmrE-dependent transport. As the researchers noted, if the protein simply fluctuated between two identical structures that differed only in membrane orientation, DEER would not distinguish conformational changes between those structures.


According to the DEER results, the unbound, apo state of EmrE is highly dynamic. This flexibility likely plays a role in the protein’s ability to bind substrates with a wide range of structures. Binding of either a proton or TPP results in a reduction in protein flexibility, which is seen most dramatically for TPP binding (Figure 2). Switching from the TPP-bound to the protonated state results in major structural changes, particularly in TM1, TM2, and TM3, and in the loops joining these TMs (Figure 3). Although the investigators carried out most of their studies in detergent micelles, similar findings were obtained when they placed EmrE in lipid bilayers. Studies of an E14Q mutant protein confirmed that most of the observed conformational changes were associated with protonation/deprotonation of glutamate-14. However, this was not true for conformational changes observed in loop 3, which appeared to be associated, instead, with protonation/deprotonation of glutamate-25 and/or aspartate-84.




FIGURE 3. Differences in the distance between residues when comparing the conformation of EmrE at pH 5 to the TPP-bound conformation. The ribbon structure on the right reflects these changes by showing increasing distance difference by increasing thickness of the ribbon. Image reproduced by permission from R. Dastvan et al., (2016) Proc. Natl. Acad. Sci. U.S.A., published online January 19, DOI:10.1073/pnas.1520431113. Copyright 2016, R. Dastvan, et al.



To better visualize the conformational changes in EmrE that were measured by the DEER studies, the Meiler lab used the data to conduct de novo modeling of the protonated conformation. In addition, the Mchaourab lab refined the existing crystal structure using DEER-defined distances to obtain a refined model of the TPP-bound conformation. The results revealed substantial flexibility in TM1, attributable, at least in part, to glycine residues at positions 8 and 9. This flexibility enables extensive rotation of the N-terminal region of TM1 upon deprotonation of glutamate-14. Similarly, deprotonation of glutamate-25 and/or aspartate-84 results in major repacking of loop 3, which translates to the terminus of TM3. These changes in the conformations of TM1 and TM3 may be responsible for the decrease in affinity for substrate that occurs with proton binding. Rotation and tilting of TM2 upon deprotonation opens a gate formed by tyrosine-40 and phenylalanine-44, providing access to the active site directly from the lipid bilayer (Figure 4A).



FIGURE 4. (A) Representation of the changes that occur in each of the TMs on the switch from the protonated to the TPP-bound conformation. (B) Structure of EmrE as depicted by molecular modeling based on the DEER results. In (a), the protein is in the resting state. It is protonated, but water is occluded. (b) TPP gains access to the protein from the inner leaflet of the membrane, promoting release of the glutamate-14-bound protons, and resulting in the TPP-bound conformation. (c) A conformational exchange of the monomers opens access to the extracellular side of the membrane. This enables release of TPP outside of the cell, followed by reprotonation of glutamate-14. Image reproduced by permission from R. Dastvan et al., (2016) Proc. Natl. Acad. Sci. U.S.A., published online January 19, DOI:10.1073/pnas.1520431113. Copyright 2016 R. Dastvan, et al.



The modeling results provided the foundation for a working hypothesis of how EmrE-dependent transport occurs (Figure 4B). The hypothesis suggests that the protonated state of the protein is a water-occluded resting state. It is important that access to the active site be restricted under these conditions, otherwise protons could easily pass through the protein, leading to dissipation of the proton gradient. Binding of substrate occurs from the lipid bilayer through the tyrosine-40/phenylalanine-44 gate. This promotes release of the protons and results in the substrate-bound state. Once substrate binding has occurred, the two subunits can undergo conformational switching, as originally hypothesized. This inverts the site of access to the active site, allowing the substrate to escape into the extracellular environment. At this point, protons may again enter the active site, allowing the cycle to repeat.


These new insights into the function of a prototypical SMR transporter are important to our understanding of how bacteria survive in the presence of some forms of toxic xenobiotics. We look forward to the further exploration of this intriguing model.



Click here to view S. Mchaourab PNAS article.









The Vanderbilt Institute of Chemical Biology, 896 Preston Building, Nashville, TN 37232-6304, phone 866.303 VICB (8422), fax 615 936 3884
Vanderbilt University is committed to principles of equal opportunity and affirmative action. Copyright © 2014 by Vanderbilt University Medical Center