One-step drug transport across two membranes of Gram-negative bacteria by an MFS tripartite assembly

One-step drug transport across two membranes of Gram-negative bacteria by an MFS tripartite assembly | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Article One-step drug transport across two membranes of Gram-negative bacteria by an MFS tripartite assembly Ben Luisi, Dijun Du, Yanjie Chao, Zhaojun Zhong, Tuerxunjiang Maimaiti, and 11 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-5901545/v1 This work is licensed under a CC BY 4.0 License Status: Published Journal Publication published 16 Mar, 2026 Read the published version in Nature Communications → Version 1 posted You are reading this latest preprint version Abstract Transport of proteins and small molecules across cellular membrane is crucial for bacterial interaction with the environment and survival against antibiotics. In Gram-negative bacteria that possess two layers of membranes, specialized macromolecular machines are required to transport substrates across the envelope, often via an indirect stepwise procedure. The major facilitator superfamily (MFS)-type tripartite efflux pumps and type I secretion systems likely employ a direct one-step transport approach, but the structural mechanisms remain elusive. A representative MFS-type tripartite efflux pump, EmrAB-TolC, mediates resistance to multiple antimicrobial drugs through proton-coupled EmrB, a member of the DHA2 transporter family. Here, we report the high-resolution (3.13 Å) structure of the EmrAB-TolC pump, revealing a unique, asymmetric architecture emerging from the assembly of TolC:EmrA:EmrB with a ratio of 3:6:1. This structure reveals two critical subdomains AssA and AssB that are essential for the pump assembly. Key residues involved in drug recognition, proton translocation and coupling are identified and corroborated by mutagenesis and antibiotic sensitivity assays. The delineation of the complete translocation pathway unravels the one-step drug transport mechanism of the MFS pump across the entire envelope of Gram-negative bacteria. Biological sciences/Structural biology/Electron microscopy/Cryoelectron microscopy Biological sciences/Microbiology/Antimicrobials/Antibiotics Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Introduction Antimicrobial resistance of Gram-negative bacteria is one of the most devastating issues in modern healthcare systems worldwide. The complex architecture of the cell envelope in Gram-negative bacteria presents a formidable obstacle for the penetration and transport of antimicrobial drugs. As one way to develop drug resistance, these organisms have evolved an array of specialized macromolecular nanomachines capable of extruding internalized drugs back to the environments 1 , 2 , for example the tripartite efflux pumps, double membrane-spanning nanomachines that are closely related with type I secretion systems (T1SSs) 1 , 2 . To cross the double membranes of Gram-negative bacteria, substrates may be exported directly from the cytosol to the outer membrane (OM) in a one-step manner, or indirectly in a two-step fashion in which substrates are first moved to the periplasm through a transporter system on the inner membrane (IM), and then translocated across the OM by another 2 . Of three classes of tripartite efflux pumps, two operate on the two-step transport mechanism 8 , including the resistance-nodulation-cell division (RND) types exemplified by AcrAB-TolC 5 , 6 and the ATP-binding cassette (ABC) types represented by MacAB-TolC 7 . The MFS-type tripartite efflux pumps are believed to employ a one-step transport mechanism like T1SSs, however the molecular details of the one-step process remain unclear 3 , 4 . The MFS-type tripartite efflux pumps are conserved in many Gram-negative bacterial pathogens 3 , 4 . The prototype MFS-type EmrAB-TolC pump of Escherichia coli plays an essential role in mediating multi-drug resistance and contributes to virulence phenotypes 9 – 13 . This pump consists of three components: a TolC OM channel, an EmrB IM transporter, and an EmrA periplasmic adaptor which bridges the two transmembrane proteins 9 . Understanding of the overall structure and organization of the EmrAB-TolC pump is currently limited to the structural data of individual, separate components of its homologues 3 , 4 . The Aquifex aeolicus EmrA structure exhibits a linear arrangement of α-helical hairpin, lipoyl, and β-barrel domains. A conserved loop in the β-barrel domain is much longer than that of other adaptors, which is disordered in the determined structure 14 . EmrB belongs to the DHA2 transporter family and, unlike the well-characterized DHA1 family for which structures in various conformations are available 15 – 18 , the DHA2 family has been less well characterized. However, recent studies on an analogous DHA2 transporter QacA have revealed critical acidic residues essential for substrate recognition and transport 19 . Structural analysis of another DHA2 transporter MHAS2168 from Mycobacterium hassiacum revealed the extension of two transmembrane (TM) helices, TM 11 and TM12, into the periplasm, which might interact with a lipoprotein for substrate translocation 20 . The stoichiometry of the EmrAB-TolC assembly remains undefined, with studies reporting conflicting models of either 3:6:1 or 3:6:2 TolC:EmrA:EmrB ratios 14 , 21 , 22 . The high-resolution structure of the full pump assembly and the mechanism of the pump are currently unknown. In this study, we present the cryo-EM structure of the E. coli EmrAB-TolC assembly for the first-time, revealing how the pump confers multi-drug resistance to antibiotics. The pump structure pinpoints several key residues that are required for drug export, mutation of which abrogate bacterial resistance to multiple antibiotics. The architecture of the EmrAB-TolC pump uncovers a one-step mechanism that directly transport antimicrobial drugs across the entire cell envelope of Gram-negative bacteria, offering new insights on the mechanisms of antimicrobial resistance. Results Engineering stable and functional EmrAB-TolC complexes Yousefian et al. demonstrated that the native tripartite EmrAB-TolC complex can be obtained, but the yield is insufficient for cryo-EM analysis 9 , because the protein complex is highly susceptible to both dissociation and precipitation. To stabilize the pump assembly, we first modified EmrB by fusing a thermostabilized apocytochrome b562RIL (BRIL) protein and an ALFA tag to the N- and C-termini of the protein 23 – 25 , respectively. We then engineered an EmrB-EmrA fusion protein with a flexible linker that connected the modified EmrB and EmrA. We observed that the N-terminus of EmrA and the C-terminus of EmrB are likely adjacent to the cytoplasmic side, and EmrA possesses a single N-terminal TM helix. Consequently, EmrA was fused to the C-terminus of modified EmrB via a flexible polyglycine-serine linker, thereby maintaining the correct membrane topology of the pump components. Co-expression of the fused EmrB-EmrA protein, free EmrA, and TolC led to the formation of a complex (engineered pump-FA). However, despite successful expression, the complex tended to dissociate or precipitate during purification, resulting in protein concentrations of only ~ 0.1 mg/mL. Therefore, the N-terminal segment (residues 1−47) of free EmrA (containing the N-terminal TM helix) was replaced by the AcrA signal sequence (residues 1−27) containing residue C25, which can be modified by a palmitoyl acid chain to anchor EmrA to the IM 26 , 27 . The modified free EmrA, EmrB-EmrA fusion protein and TolC were successfully expressed and copurified, and the complex (engineered pump-EA) exhibited modest stability during purification, achieving a concentration of ~ 0.7 mg/ml. The genomic emrAB is expressed poorly in E. coli under laboratory growth conditions, and its knockout does not significantly affect resistance 28 . However, mutations that increase the expression of the pump lead to increased resistance 12 . Therefore, we induced the expression of the pump to carry out survival assays, which revealed that the wild-type EmrAB-TolC pump conferred drug resistance to nitroxoline, nalidixic acid and CCCP (Extended Data Fig. 1 a). Induced expression of the engineered EmrAB-TolC pump-EA pump showed that it conferred reduced resistance to nitroxoline, indicating that the complex was functional in vivo . The EmrAB-TolC pump-FA, however, exhibits no resistance activity (Extended Data Fig. 1 b). Quaternary structure of the EmrAB-TolC pump The structure of EmrAB-TolC was solved using single-particle cryo-EM. Three maps were reconstructed: two maps for pump-EA achieved overall resolutions of 3.13 Å and 3.14 Å, respectively; and one map for pump-FA reached 3.59 Å resolution (Extended Data Fig. 2 a-c; Extended Data Fig. 3 a-c). Local resolution examination of these maps indicated a resolution ranging between 3.0 Å to 7.0 Å (Extended Data Fig. 2 d; Extended Data Fig. 3 d). The map quality of EmrAB-TolC pump-FA was insufficient for reliably constructing the full pump structure, whereas the maps for pump-EA were of sufficient quality to allow building models for residues 47−390 of EmrA and residues 13−498 of EmrB (Fig. 1 ; Extended Data Fig. 2 e). Nevertheless, the model derived from the EmrAB-TolC pump-EA map fits well within the EmrAB-TolC pump-FA map. The models built from both maps show minimal differences, indicating that the pump-EA variant does not disrupt the assembly or structure of the complex. The map for pump-FA was used to construct models for the extra residues 17−46 of EmrA (Fig. 2 a-c; Extended Data Fig. 4 ). The pump has an elongated shape consisting of a TolC trimer, an EmrA hexamer and an EmrB monomer (Fig. 1 ; Extended Data Fig. 4 ). In the case of pump-FA, the model of the N-terminal TM helix of the EmrA protomer 4 spans the entire IM, while the partial models of the other five EmrA protomers are shorter (Extended Data Fig. 4 ; Extended Data Fig. 5 ). The positions of EmrB and TolC delineate the IM and OM boundaries, respectively, with the long axis of the pump assembly extending ~ 320 Å into the periplasm (Fig. 1 ; Extended Data Fig. 4 ). This is comparable to the dimensions observed for other types of tripartite assemblies, such as AcrAB-TolC and MacAB-TolC 5 – 7 . The EmrA hexamer forms a nanochannel with an AssA subdomain for pump assembly The cryo-EM map revealed five sequentially ordered structural units in EmrA: the singular N-terminal TM helix and the α-helical hairpin, lipoyl, and β-barrel domains, as well as a C-terminal α-helix (Fig. 2 a-c; Extended Data Fig. 5 ). The N-terminal TM helix anchors the four periplasmic units to the IM (Fig. 2 b, c). Adaptor proteins, such as AcrA and MacA, utilize a MP domain for interaction with the periplasmic extension of IM transporters 5 – 7 . EmrA and its homologues, as well as the adaptor proteins of T1SS, however, lack such a MP domain (Extended Data Fig. 6 ; Extended Data Fig. 7). The α-helical hairpin domain structure resembles that of A. aeolicus EmrA, although it is one-third shorter (Extended Data Fig. 6 a, b) 14 . The structures of both the lipoyl and β-barrel domains are similar to those of other adaptors (Extended Data Fig. 6 ; Extended Data Fig. 7). A notable difference is the presence of a unique loop (termed β-CL), comprising residues 294−317, in the β-barrel domain of EmrA (Fig. 2 b; Extended Data Fig. 8a), which is conserved among EmrA homologues (Extended Data Fig. 9a). A long coil follows the C-terminus of the β-barrel domain, folding back to engage with the side of the adjacent lipoyl domain in the pump assembly. This coil precedes an α-helix at the terminus (residues 375−390) that, together with the lower portion of the α-helical hairpin, forms a three-stranded coiled-coil, mimicking the α-helical hairpin domain of CusB 29 . The α-helical hairpin, lipoyl and β-barrel domains of EmrA assemble into a hexameric structure reminiscent of AcrA and MacA within the MacAB-TolC and AcrAB-TolC pumps (Fig. 2 c; Extended Data Fig. 7) 5 , 7 . The helical hairpin domains create an α-helical barrel nanochannel, referred to as NanoHP. The lipoyl domains form a ring encompassing an internal chamber that extends the nanochannel (named CavLP). The analogous hexameric assembly of MacA presents a gating ring formed by conserved glutamine residues in the lipoyl domain 7 , but this is notably absent in the EmrA assembly. The six β-CLs within the β-barrel domains (β-CL1 to β-CL6) vary in conformation, creating a structural unit which we refer to as assembly subdomain A (AssA), which forms the wall of a chamber (named CavEA) (Extended Data Fig. 8a-c). Residues L302 and L303 in the β-CL hexamer form a hydrophobic ring-1 on the IM distal side of CavEA, providing a narrow entryway to CavLP. Conversely, residues I313, V315 and V316 generate a hydrophobic ring, ring 2, on the IM proximal side of CavEA, potentially situated on the surface of the cellular membrane (Extended Data Fig. 8b, c; Extended Data Fig. 9a). The β-CL1 contains an alternative route (termed β-CL1PH) linking the chambers CavEA and CavLP (Extended Data Fig. 8c). A monomer of EmrB containing an AssB subdomain mediates pump assembly EmrB, which belongs to the DHA2 family, comprises 14 transmembrane helices, of which H1−H12 adopt the characteristic fold of MFS transporters such as MdfA and EmrD 16 , 17 . The six-helix bundles, formed by H1−H6 (termed EBN) and H7−H12 (termed EBC), create a V-shaped transporter (Fig. 3 a-c) with a pseudo-twofold symmetry axis perpendicular to the membrane plane. Two more TM helices, H A and H B , are inserted into the cytoplasmic loop bridging the EBN and EBC domains, but against the periphery of the protein core, similar to YbgH and PepTso 30 – 32 . The H A and H B counterparts of QacA and Tet(L) are known to be important for their functions 19 , 33 . Structural analysis of EmrB revealed a central aqueous cavity, CavEB out , located in the core of the membrane that opens toward the periplasmic space, indicating that EmrB adopts an outward-open state (Extended Data Fig. 10a). Entry into the cytoplasm from this cavity is restricted by interactions of side chains between helices H4 and H5 in the EBN domain and helices H10 and H11 in the EBC domain (Extended Data Fig. 10a). Within the central cavity, we identified a proton-titratable residue D29, located in motif D, which we hypothesized would be critical in the antiporter function of EmrB 34 . The corresponding residues of D34 in the TM1 of QacA and MfdA have previously been shown to be crucial for substrate recognition 17 , 18 . In our structure, D29 forms a salt bridge with R109 and hydrogen bonds with Q112 and N174, respectively (Fig. 3 d). All three of these residues are highly conserved among EmrB homologues (Extended Data Fig. 9b). The EmrB D29N or EmrB R109A mutation was found to abrogate bacterial resistance to nalidixic acid, nitroxoline and CCCP, supporting the hypothesis that these residues are involved in proton translocation during transport (Fig. 3 e). Interestingly, EmrA or EmrB alone is inactive, suggesting that the EmrAB complex is crucial for its function (Fig. 3 e). Parts of the 13th and 14th TM helices (H11 and H12) of EmrB protrude into the periplasm and form a structural unit named assembly subdomain B (AssB) (Fig. 3 a-c). The AssB is composed of an N-terminal helix X (H X ) following H11, a C-terminal helix Z (H Z ) preceding H12, and a horizontal helix Y (H Y ) parallel to the membrane, along with the loops connecting these helices. The most pronounced difference between EmrB and other MFS transporters is the unique AssB subdomain structure, which is key for pump assembly (Fig. 3 b). Interfacial contacts between EmrA and EmrB Interactions of EmrB with other components of the tripartite assembly appear to be required for efflux activity, suggesting that the interactions allosterically modulate the activity of the MFS protein. The β-barrel domain of EmrA mediates interaction with EmrB, with the AssA subdomain of EmrA playing a key role through interactions with the periplasmic AssB subdomain of EmrB (Fig. 4 a). H Y in the AssB subdomain penetrates the hydrophobic ring-1 in the AssA subdomain. Since H Y is larger than the narrow entrance formed by hydrophobic ring-1, it becomes trapped above the hydrophobic ring-1, effectively blocking the narrow entrance (Fig. 4 b, c; Extended Data Fig. 8b, c). The EmrA L302Q/L303Q mutation abrogated pump-mediated resistance to nalidixic acid, nitroxoline, and CCCP, indicating that the hydrophobic ring-1 plays an essential role in pump assembly (Fig. 4 d, e; Extended Data Fig. 8c). Below the hydrophobic ring-1, helices H X and H Z in the AssB subdomain tilt towards one side of CavEA, making contacts only with β-CL1 to β-CL3 (Fig. 4 b). In addition, the cytoplasmic surface of the EmrA β-barrel domains directly engages with the periplasmic region of EmrB, and the N-terminal TM helices of EmrA insert into the membrane, stabilizing the interaction interface. The protruding loops in EmrB, positioned between helices H1−H2, H7−H8 and H9−H10, fit snugly into the concave pockets in the AssA subdomain. Helices H X and H Z of the AssB subdomain, along with other periplasmic regions between the transmembrane helices of EmrB, position along the side of the CavEA. The internal chamber formed is referred to as CavEA out , and its size is sufficient to accommodate all known EmrB substrates. The cavity CavEB out in EmrB opens to CavEA out , allowing the substrate in CavEB out to release into CavEA out (Fig. 4 c). Given the blockage of the narrow entrance by H Y , the route of β-CL1PH in β-CL1 serves as the sole link between CavEA out and CavLP. The hydrophobic ring-2 in AssA assists in the formation of a sealed continuous channel between CavEB out and CavEA out by resting on the surface of the cellular membrane (Extended Data Fig. 8b, c). The EmrA I313Q/V315T/V316T mutation in hydrophobic ring-2 abrogated resistance to nalidixic acid, nitroxoline and CCCP, indicating that the hydrophobic ring-2 is critical for the pump assembly (Fig. 4 d, e; Extended Data Fig. 8c). Prediction of EmrB in its inward-open conformation EmrB has been suggested to adopt an inward-open state for substrate recognition and binding during transport, thereby forming a periplasmic gate 35 . However, in both our cryo-EM structure and all 5 models predicted with high-confidence scores by AlphaFold3, EmrB is consistently displayed in an outward-open state (Fig. 5 a; Extended Data Fig. 11a, b) 36 . In an attempt to capture EmrB in its inward-open state, we utilised the multimer feature of AlphaFold3 to generate a prediction of EmrB in complex with six molecules of EmrA. In all 5 of the models generated by AlphaFold, we were able to observe EmrB in a distinct conformation (Fig. 5 b; Extended Data Fig. 11c, d). Based on the disappearance of CavEB out and the presence of a new and enlarged cytosolic cavity (named CavEB in ), we hypothesized that AlphaFold had been able to capture the inward-open state of EmrB previously unobservable in the experimental data set (Fig. 5 c) 35 . Using a homology model of EmrB in the inward-open state in conjunction with these predictions, we identified several residues which we believed would be important in the process of translocation. Residues N62, V147, I148 and I288 are all situated along the wall of the cytosolic-facing cavity (Extended Data Fig. 10b). The corresponding residues in MdfA are involved in drug recognition (Extended Data Fig. 12) 18 . In support of AlphaFold, mutations of N62, V147, and I288 indeed reduced pump-mediated resistance to nitroxoline and nalidixic acid. I148 was crucial for resistance to CCCP, whereas it appeared dispensable for nitroxoline and nalidixic acid (Fig. 5 f, g). Nalidixic acid/H⁺ antiport activity was detected in everted membrane vesicles expressing wild-type EmrB and its mutants. Wild-type EmrB exhibited significant H⁺ movement, whereas the EmrB D29N and EmrB V147A mutants showed reduced H⁺ movement (Fig. 5 h). Altogether, these results support the predicted model for the inward-open state and demonstrating the importance of specific residues within EmrB in exhibiting poly-specific drug recognition during the transport process. Proposed structural transitions of EmrB Upon close inspection of the structures of EmrB in different states, we detected significant changes in the chemical environments of residues already revealed to be important in our mutagenesis studies. For example, the proposed proton-carrier D29 was shown to establish 3 hydrogen bonds in the outward-open state in comparison to just 2 hydrogen bonds in the inward-open state (Fig. 5 d). To provide further evidence that AlphaFold3 had indeed captured the inward-open state and to try gain insight into conformational changes within EmrB, we employed a modified version of AlphaFold2, known as AF-Cluster, to obtain predictions of the various states which EmrB could adopt 37 . By grouping the multiple-sequence alignments based on sequence similarity within the MFS family and subsequently performing AlphaFold2 predictions, we captured a plethora of structures representative of the transition between the outward-open to inward-open state of EmrB (Fig. 5 e). Interestingly, 75% of these high-confidence predictions could be classified into structures representative of the inward-open states, providing further evidence of this conformation being adopted by EmrB. Furthermore, these predictions also provided insight into the conformational dynamics that EmrB is likely to undergo in shuttling substrates, analogous to the transitions proposed for other members of the MFS family 35 . The helices within the EBC domain appear to remain static, whereas helices H1 – H6 in the EBN domain are suspected to be highly dynamic. The helices appear to pivot at a central point, in support of a ‘rocker-switch’ mechanism of transport for EmrB, common to other MFS antiporters 38 . Altogether, these results provide evidence of an inward-open conformation of EmrB, as well as insights into the structural transitions of EmrB that are important in the energy-dependent flux of antibiotics by EmrAB-TolC. The interactions of EmrA-EmrB change in the structural transitions of EmrB, and may provide the path of allosteric communication of these components. Interactions between EmrA and TolC A short helix-turn-helix motif located within the α-helical hairpin domain of EmrA mediates tip-to‐tip interactions with the periplasmic end of TolC, similar to the interactions observed between the homologous MacA and TolC (Extended Data Fig. 13) 7 . Engagement between TolC and EmrA directly opens the periplasmic aperture of TolC, creating a continuous nanochannel with a width of 25−30 Å, termed NanoTC. TolC is predicted to remain open during the pumping process, as observed previously for RND-type AcrAB-TolC and MexAB-OprM pumps 5 , 7 , 39 . One-step drug transport mechanism of the EmrAB-TolC pump The structure of the EmrAB-TolC tripartite pump represents the resting apo state structure. The combined data with the Alphafold3 prediction structures suggest that it employs a one-step drug transport mechanism across the entire cell envelope (Fig. 6 ). EmrA and EmrB possess unique AssA and AssB subdomains, respectively (Fig. 3 b; Extended Data Fig. 8b, c). The six β-CLs in AssA vary in conformation, one of which develops a path (β-CL1PH) linking the chambers CavLP and CavEA out in the lipoyl and β-barrel domains (Fig. 4 b, c; Extended Data Fig. 8a-c). Residues L302 and L303 in the β-CL hexamer form a hydrophobic ring-1, onto which the AssB anchors (Fig. 4 ; Extended Data Fig. 8b, c). EmrA lacks the MP domain, which enables EmrA and EmrB to form a secure conduit and its hydrophobic ring-2 strengthens this conduit (Fig. 4 d; Extended Data Fig. 8b, c). The structure of the EmrAB-TolC pump delineates the complete transport pathway from the central cavity CavEB out of EmrB to the cell exterior (Extended Data Fig. 14). As a homologue of QacA and MdfA 17 , 18 , 40 , EmrB in the IM could adopt an inward-open state to facilitate substrate recognition and binding from the inner leaflet of the IM and cytoplasm, as supported by AlphaFold3 predictions, homology modelling and mutagenesis studies (Fig. 5 f; Fig. 6 ; Extended Data Fig. 10b). Subsequently, the substrate is transported to CavEA out via an alternating access mechanism, characteristic of MFS transporters (Fig. 6 ) 35 . The protonation of D29 transforms EmrB to an inward-open state again. The periplasmic regions between the transmembrane helices of EmrB shift toward the central axis, which partially occupy the space in CavEA and cause CavEA out to shrink. The compaction of CavEA out drives substrates toward the β-CL1PH-CavLP pathway to the nanochannel formed by EmrA and TolC. This contraction of CavEA has been proposed in the AlphaFold3-predicted EmrAB structure (Fig. 5 c). The transport pathway involves several essential elements, namely, CavEB in , CavEB out , CavEA out , β-CL1PH, CavLP, NanoHP, and NanoTC (Extended Data Fig. 10a, b; Extended Data Fig. 14). These components form a secure conduit for substrate release to the cell exterior (Fig. 6 ). The gating helices in EmrB allow substrates to flow against concentration gradients and prevent backflow during operation (Extended Data Fig. 10a, b), resulting in the export of drugs across the IM and OM to the extracellular milieu. Discussion Antibiotic resistance of pathogenic bacteria is a growing clinical problem. Drug efflux pumps in the bacterial cell envelope play important roles in multidrug resistance. In the current study, the E. coli EmrAB-TolC pump was shown to confer resistance to several different antimicrobial drugs, including nalidixic acid, nitroxoline and CCCP (Extended Data Fig. 1 a). The EmrAB-TolC structure identifies key residues that contribute to such multi-drug resistance (Fig. 4 d; Fig. 5 f). The conserved hydrophobic ring-1 of EmrA is crucial for the assembly and functioning of the pump and is accessible to the external environment via the open nanochannel in the full pump assembly (Extended Data Fig. 9a; Extended Data Fig. 14); thus, it represents a promising target for the development of drugs that inhibit this class of pump. At present, six families of efflux pumps are known to be involved in drug transport: primary transporters of the ATP-binding cassette (ABC) family, and secondary transporters including the major facilitator (MFS), resistance/nodulation/cell division (RND), small multidrug resistance (SMR), multidrug and toxic compounds extrusion (MATE), and proteobacterial antimicrobial compound efflux (PACE) families. In Gram-negative bacteria, all these drug transporters are located in the IM. The members of five families of drug transporters (MFS, SMR, MATE, PACE and ABC) usually function as independent units to translocate substrates across the IM bilayer 8 . Drugs on the periplasmic side must be further translocated across the OM. Structures of RND-type AcrAB-TolC and ABC-type MacAB-TolC reveal that the IM components AcrB and MacB collect substrates from the outer leaflet of the IM and periplasm, and efflux them to the extracellular milieu through the tripartite assembly (Extended Data Fig. 15a, b) 5 – 7 . It is likely that the five transporter systems cooperate with the double membrane-spanning RND-type or ABC-type tripartite pumps to deliver substrates across the OM to the cell exterior. Thus, these tripartite efflux pumps employ a two-step mechanism of periplasmic substrate capture and transport (Extended Data Fig. 15a, b) 2 , 8 . On the other hand, MFS-type tripartite pumps appear to utilize a one-step mechanism to transport drugs across the entire cell envelope, bypassing the periplasm (Extended Data Fig. 15c). The closely related T1SS assemblies, such as HlyDB-TolC, are believed to employ similar processes for the secretion of protein substrates 1 . The HlyDB-TolC pump could form a tripartite assembly. Similar to EmrA, the adapter protein HlyD lacks a MP domain (Extended Data Fig. 6 ) 4 , 41 , which may enable HlyD and HlyB to form a secure conduit. The ABC family transporter HlyB could drive the secretion substrate HlyA across double membranes of the cells via an alternating access mechanism. We note that RND-type tripartite pumps can employ a one-step transport mechanism across the outer membrane for certain substrates, such as beta-lactam antibiotics 42 , but it is rather different mechanistically from the EmrAB-TolC pump. The EmrAB-TolC, AcrAB-TolC and MacAB-TolC pumps all utilize the TolC OM channel, and have structurally similar periplasmic adaptor proteins that form hexametric channels, but their IM components are derived from distinct transporter families. The periplasmic adaptor proteins have evolved uniquely to complement the specific IM transporters. As a result, each type of tripartite pump utilizes a unique pump assembly and gating mechanism that contributes to its function. The ability of these proteins to recognize and transport various substrates largely depends on the specific domains within their IM transporter components, and interactions of the components can allosterically modulate the activity of the inner membrane component. Here, we observe extensive interactions between EmrA and EmrB that change with conformational switch. These interactions are proposed to modulate the activity of EmrB so that it does not function as a stand-alone transporter, as occurs for example in the well-studied MFS proteins QacA and MfdA. Understanding the similarities and differences between tripartite multidrug efflux pumps and the basis for their allostery is critical for developing strategies to combat multidrug resistance, a global health challenge. Materials And Methods Construction of EmrAB-TolC complex overexpression vectors The BRIL gene was synthesized by AZENTA (Suzhou, China) and amplified using the brilpet-F and bril-link-R primers. The emrA , emrB , tolC , and acrA genes were amplified from genomic DNA of E. coli K-12 strain MG1655. The emrB (37-1494 bp) genefragmentwas amplified using primers emrB-link-F and emrBGS-R. The full-length emrA gene was first amplified using primers emrAGS-F and emrApet-R; the pETDuet-1 plasmid was linearized using primer pair Petduet_F and Petduet_R; and the BRIL , emrB , and emrA DNA fragments were inserted into the pETDuet-1 vector using NovoRec plus One step PCR Cloning Kit (Novorprotein, Shanghai, China), resulting in pETDuet- BRIL-emrB-polyGlySer-emrA-6His . The ALFA-tag sequence was subsequently inserted into the 3′ end of emrB by site-directed mutagenesis, generating pETDuet- BRIL-emrB-ALFA-polyGlySer-emrA-6His (abbreviated as pET_BA). Using primers TolCinf_F and TolCFLAGXhoI_R, the tolC gene was amplified, and the product served as a template for a second amplification using primers TolCinf_F and TolCFLAG_inf_R. The resulting tolC-FLAG DNA fragment was inserted into pRSFDuet-1 (digested with Nde I) using the in-fusion ligation method of the NovoRec plus One step PCR Cloning Kit (Novorprotein, Shanghai, China), thereby generating pRSFDuet- tolC-FLAG . The emrA DNA fragment was amplified using primers emrAinsert_F and emrAinsert_R and inserted into the pRSFDuet- tolC-FLAG vector using the in-fusion cloning method (Novorprotein, Shanghai, China), resulting in pRSFDuet- emrA-tolC-FLAG (abbreviated as pRSF_AC). Primer pair Δ emrAinsert_F and emrAinsert_R was used to amplify the truncated ΔemrA gene (142-1173 bp), and the ΔemrA DNA fragment was inserted into the pRSFDuet- tolC-FLAG vector using the In-Fusion cloning method (Novorprotien, Shanghai, China), generating pRSFDuet- ΔemrA-tolC-FLAG . The N-terminal 81 bp region of acrA ( acrAsignal ) was amplified using the acrAs_F and acrAs_R primer pair, and the resulting acrAsignal DNA fragment was inserted into the 5' end of ΔemrA in pRSFDuet- ΔemrA-tolC-FLAG using the In-fusion cloning method (Novorprotien, Shanghai, China), generating pRSFDuet- acrAsignal-ΔemrA-tolC-FLAG (abbreviated as pRSF_ Δ AC). The emrAB genes were amplified from genomic DNA of E. coli K12 strain MG1655 using primers EmrAB_F and EmrAB_R, the pETDuet-1 plasmid was amplified using primers Petduet_V_F and Petduet_V_R, and the emrAB fragment was inserted into pETDuet-1 using an In-Fusion Ligation Kit, resulting in pETDuet- emrAB . The site-directed mutants EmrAB D29N and EmrAB R109A were produced using pETDuet- emrAB as a template with the primer pairs emrB D29N _F/emrB D29N _R and emrB R109A _F/emrB R109A _R, resulting in pETDuet- emrAB D29N and pETDuet- emrAB R109A . The site-directed mutants EmrAB N62A , EmrAB V147A , EmrAB V148A , and EmrAB I288A were created by using pETDuet- emrAB as a template with the primer pairs emrB N62A _F/emrB N62A _R, emrB V147A _F/emrB V147A _R, emrB V148A _F/emrB V148A _R, and emrB I288A _F/emrB I288A _R, resulting in pETDuet- emrAB N62A , pETDuet- emrAB V147A , pETDuet- emrAB V148A , and pETDuet- emrAB I288A . The site-directed mutants EmrA L302Q/L303Q B and EmrA I313Q/V315T/V316T B were produced using pETDuet- emrAB as a template with the primer pairs emrA L302Q/L303Q _F/emrA L302Q/L303Q _R and emrA I313Q/V315T/V316T _F/ emrA I313Q/V315T/V316T _R, resulting in plasmids pETDuet- emrA L302Q/L303Q B and pETDuet- emrA I313Q/V315T/V316T B . The enzyme used to introduce the mutations above-mentioned was purchased from CloneAmp HiFi PCR Premix (Clontech, Germany). Then the modified plasmids above-mentioned were amplified and extracted using SPARKeasy Mini Plasmid Ultra-Fast Kit (Shandong Sparkjade Biotechnology Co.,Ltd.). Primers used to generate all constructs are listed in Extended Data Table 1. Overexpression and purification of E. coli EmrAB-TolC The plasmids pET_BA and pRSF_ Δ AC were used to co-express and purify EmrAB-TolC pump-EA, which contains the EmrB-EmrA fusion protein, TolC and a modified free EmrA with the N-terminal residues 1-47 replaced by residues 1-27 of AcrA. The 6xHis-Tag at the C-terminus of the EmrB-EmrA fusion protein was initially used to pull down the co-expressed pump components. Next, the FLAG-Tag at the C-terminus of TolC was used to isolate the fully assembled EmrAB-TolC pump. BRIL and ALFA-Tag was used to stabilize EmrB in the complex. The E. coli C43 (DE3) ΔacrAB strain was transformed with plasmids pET_BA and pRSF_ Δ AC. From an agar plate containing appropriate antibiotics, a single colony was picked and inoculated into 20 mL of LB medium supplemented with 50 μg mL -1 kanamycin and 100 μg mL -1 carbenicillin in a 50 mL centrifuge tube. The culture was incubated at 37°C with shaking at 220 rpm for 4-5 h. A 10 mL culture aliquot was then transferred to 1 L of 2´YT medium containing antibiotics in a 2 L baffled flask. The same incubation conditions were applied. Cells were cultured until a density (absorbance at 600 nm, A 600 ) of 0.5-0.6, induced with 0.1 mM isopropyl-β-D-thiogalactoside (IPTG), the temperature was reduced to 20°C, and culturing was continued overnight. The cells were harvested by centrifugation, and the cell pellet from 6 L of culture was resuspended in 200 mL of lysis buffer containing 20 mM Tris (pH 8.0) and 300 mM NaCl. Afterwards, 2 mL of 100´EDTA-Free Protease Inhibitor Cocktail (APExBIO, USA) was added to a final concentration of 1´. Lysozyme and DNase I were added to final concentrations of 5 mg/mL and 5 U/mL, respectively. The cell mixture was incubated at 4°C for 1 h before high-pressure homogenization via three passages at 15,000 psi at 4°C. The lysate was centrifuged to remove cell debris, and the supernatant was ultracentrifuged to pellet the cellular membranes. Lysis buffer was used to resuspend the membrane pellet. EDTA-free protease inhibitor cocktail (100´) was added to a final concentration of 1´, n -dodecyl- β -D-maltoside (DDM) was added to a final concentration of 1.5%, and the mixture was gently stirred at 4°C for 3 h. The mixture was ultracentrifuged. Imidazole was added to the supernatant to a final concentration of 20 mM, and the mixture was applied to a 1 mL HiTrap Chelating column (Cytiva, USA) charged with Ni 2+. The column was washed with lysis buffer containing 0.05% DDM and 50 mM imidazole, followed by lysis buffer supplemented with 0.01% DDM. Peptidisc 43,44 (1 mL) at a concentration of 5 mg/mL in lysis buffer was injected onto the column and incubated at 4°C for 1 h. The column was washed with 10 mL of lysis buffer containing 1 mg/mL peptide disc. Elution of the 6´His-tagged EmrAB-TolC protein complex was achieved with lysis buffer containing 300 mM imidazole. Buffer exchange of the eluate into Buffer-I, which consisted of 20 mM Tris (pH 7.5) and 150 mM NaCl, was conducted using a HiTrap Desalting column (Cytiva, USA). The 6´His-tagged EmrAB-TolC complex was further purified using ANTI-FLAG M2 affinity resin (GenScript, Nanjing, China). The resin was prepared by sequential washing with glycine HCl (pH 3.5) and Buffer-I. The protein solution and resin were mixed, gently rotated for 1 h at 4°C, loaded into a chromatography column, and washed with Buffer-I. The mixture was resuspended in 1 mg/mL FLAG-peptide in Buffer-I, incubated for 30 min, centrifuged, and the supernatant was loaded onto a mini chromatography column to remove residual resin. This step was repeated three times. Fractions containing EmrAB-TolC were pooled, concentrated using a centrifugal filter unit (Merck, Germany) with a molecular weight cut-off of 100 kDa, flash-frozen in liquid nitrogen and stored at -80°C. The plasmids pET_BA and pRSF_AC were used to co-express and purify the EmrAB-TolC pump-FA containing the EmrB-EmrA fusion protein, TolC, and wild-type free EmrA using the same procedure employed for the pump-EA. Cell growth for the antibiotic sensitivity assay Single colonies of E. coli C43(DE3) ΔacrAB harboring the plasmids pETDuet/pRSFDuet, pET_BA/pRSF_AC, pET_BA/pRSF_ Δ AC, pETDuet- emrAB /pRSFDuet- tolC-FLAG , pETDuet- emrAB D29N /pRSFDuet- tolC-FLAG , pETDuet- emrAB R109A /pRSFDuet- tolC-FLAG , pETDuet- emrA L30Q/L303Q B /pRSFDuet- tolC-FLAG , pETDuet- emrA I313A/V315A/V316A B /pRSFDuet- tolC-FLAG , pETDuet- emrA B N62A /pRSFDuet- tolC-FLAG , pETDuet- emrA B V147A /pRSFDuet- tolC-FLAG , pETDuet- emrA B V148A /pRSFDuet- tolC-FLAG , or pETDuet- emrA B I288A /pRSFDuet- tolC-FLAG were grown in fresh LB supplemented with 50 µg/mL kanamycin and 100 µg/mL ampicillin at 37℃ with shaking at 220 rpm. When the optical density at 600 nm (OD 600 ) reached 0.5–0.6, expression was induced with 100 µM IPTG, the temperature was reduced to 22°C, and the incubation was continued for 1 h. The cells harboring pET_BA/pRSF_AC or pET_BA/pRSF_ Δ AC were harvested by centrifugation, after which the cell pellets were resuspended in fresh medium to an OD 600 of 0.1 and treated with 0.8 mg/mL nitroxoline and 30 µM IPTG. Cell growth was monitored over time by measuring the OD 600 at 37°C in a Spark microplate reader (Tecan, Austria). The cells harboring the remaining constructs were harvested by centrifugation, resuspended in fresh medium to an OD 600 of 0.5, and 10-fold serially diluted. Then 5 µL of each culture was plated on LB agar containing 60 µM IPTG in the presence of 1 µg/mL nalidixic acid, 5 µg/mL nitroxoline, or 1 µg/mL CCCP. As a control, 5 μL of all suspensions were plated on LB agar without antibiotics. Cell growth was monitored after culture at 37°C overnight. The assay was performed three times independently. Drug-proton antiport assay The assay was conducted as described previously 40,45-47 . Inside-out (everted) membrane vesicles were prepared from C43 (DE3) ΔacrAB cells expressing the EmrB variants. Briefly, bacterial cells were cultured in 250 ml LB medium at 37°C to an OD600 of 0.5, then induced for protein expression by adding 0.2 mM IPTG at 37°C for 2 hours to an OD600 of 1.5. The cells were harvested by centrifugation at 7,500 x g for 10 minutes, resuspended in 20 ml of membrane vesicle buffer A (200 mM Tris-HCl, pH 8.0, 2 mM EDTA, 30% sucrose) at room temperature, and incubated with 100 μg/ml lysozyme for 5-10 minutes. After adding 4 mM MgCl 2 and DNase I, the cell spheroplasts were pelleted at 7,500 x g for 10 minutes, then resuspended in 25 ml of ice-cold membrane vesicle buffer B (100 mM HEPES-KOH, pH 7.5, 100 mM KCl, 10 mM MgCl 2 ), supplemented with 2 mM DTT and 250 μL protease inhibitor (K1007; APExBIO). The spheroplasts were lysed by five passages through a homogenizer (ATS, AH-NANO) at 400-500 bar. The unbroken cells and debris were pelleted at 7,500 x g for 10 minutes, and the membranes in the supernatant were pelleted at 150,000 x g for 40 minutes at 4°C in a Ti70 rotor (Beckman Coulter). The membrane pellet was resuspended in membrane vesicle buffer B by gently squirting the buffer toward the pellet with a 200 μL tip. The vesicles were then centrifuged at 3,500g at 4 °C for 30 s to remove the aggregates. To ensure equal quantities of everted membrane vesicles were used for each measurement, vesicles containing 200 μg total membrane proteins were added to 2 ml of pre-warmed (30°C) buffer （20 mM Tris, 140 mM KCl, 10 mM MgCl 2 , pH 6.0） and 1 μM 9-Amino-6-chloro-2-methoxyacridine (ACMA). Membranes were equilibrated for 4 minutes prior to fluorescence measurement with excitation at 409 nm and emission at 474 nm using an FluoroMax-4 spectrofluorometer (HORIBA Scientific). Before adding the substrates, 2 mM lactate was used and equilibrated for 3 minutes to energize the membrane, quenching the ACMA fluorescence. Fluorescence dequenching was observed upon adding 250 μM nalidixic acid, likely due to the extrusion of H + by antiporters translocating the drugs into the vesicles. To dissipate the transmembrane H + gradient, 5 mM NH 4 Cl was added. The fluorescence changes induced by lactate and NH 4 Cl further confirmed that similar amounts of everted vesicles of comparable quality were used in each measurement. 6xHis-tag was added to the N-terminus of EmrA or C-terminus of EmrB of EmrAB-TolC pump, respectively. The same procedure was used to prepare everted membrane vesicles for Western blot assays of EmrA/EmrB variants (Fig. 4e; Fig. 5g). Electron microscopy data collection For the EmrAB-TolC pump-EA cryo-EM assays, 3.5 μL aliquots of purified protein in peptidisc (protein concentration = 0.7 mg/mL) were added to glow-discharged holey carbon grids (Quantifoil Au R1.2/1.3, 300 mesh; Quantifoil Micro Tools GmbH). Blotting was performed with filter paper for 3.5 s to remove excess sample, and a Vitrobot Mark IV instrument (Thermo Fisher Scientific) was used for rapid freezing in a liquid ethane slush. A Titan Krios electron microscope (Thermo Fisher Scientific) operating at 300 kV coupled with a SerialEM and a Gatan K3-Summit detector (Gatan, Inc.) operating in super-resolution counting mode 48 were used to automatically collect zero-energy-loss images of frozen and hydrated grids. Using a slit width of 20 eV, a GIF-Quantum energy filter (Gatan) was applied to exclude inelastically scattered electrons. Using a dose rate of ∼15.15 electrons Å -2 s -1 (~18 electrons pixel -1 s -1 ) at an adjusted magnification of ´45,871.6 (yielding a pixel size of 1.09 Å at the sample level) and a total dose of ∼60 electrons Å −2 at the sample, 60 movie frames were recorded. The final dataset comprised 4334 movie stacks with defocus values between -1.0 and -2.5 μm. For the EmrAB-TolC pump-FA, a holey carbon grid (Quantifoil Au R1.2/1.3, 300 mesh) was overlayed with a homemade graphene monolayer and cleaned with UV/ozone at room temperature for 10 min using a Gatan SOLARUS (950) Plasma Cleaning System (Gatan, Inc.), ensuring hydrophilicity of the graphene grid. Next, 3.5 μL aliquots of the purified protein sample from the peptidisc (0.1 mg/mL concentration) were applied to the grid and incubated for 30 s. Excess sample was removed by blotting with filter paper for 3.5 s, followed by rapid freezing in liquid ethane slush using a Vitrobot Mark IV instrument. Frozen-hydrated EmrAB-TolC particles were subjected to automatic data collection using a Titan Krios electron microscope at 300 kV with SerialEM and a Gatan K3-Summit direct electron detector running in super-resolution counting mode at an adjusted magnification of 47,169.9, equating to a measured physical pixel size of 1.06 Å and a dose rate of ∼16 electrons Å -2 s -1 (∼18 electrons pixel -1 s -1 ). Exposures lasting 3.533 s were split into 60 movie frames, resulting in an accumulated dose of ∼56.6 electrons Å -2 at the sample. Using a defocus range of -1.0 to -2.5 μm, 3306 movie stacks were collected. Image processing For the EmrAB-TolC pump-EA, the super-resolution movie frames were adjusted, including correction for gain reference and 2´ binning, followed by motion correction using MOTIONCORR2 49 . Merging of the aligned movie frames into micrographs then allowed estimation of the contrast transfer function (CTF) using CTFFIND4 50 . RELION v3.1.3 software was used for subsequent image processing steps 51,52 . Templates for automatic particle picking were derived from (reference-free) two-dimensional classification of a manually selected particle subset. To minimize reference bias, these templates underwent low-pass filtering to 20 Å, enabling the automatic selection of 816,894 particles from all the micrographs. Two-dimensional classification of the images yielded 480,902 suitable particles. Additional image processing steps were carried out using CryoSPARC 53 . Particles chosen via two-dimensional classification within RELION were imported into CryoSPARC and subjected to two further rounds of two-dimensional classification to discard any obviously discrepant particles, yielding 56,006 particles for classification by ab initio reconstruction using CryoSPARC. The parameters for this classification included two-class ab initio reconstruction with specific settings (initial alignment resolution of 25 Å; maximum alignment resolution of 6 Å; initial minibatch size of 150; final size of 600; and class similarity of 0) 54 . A total of 33,460 particles contributed to the resulting 3D volume, which was subjected to nonuniform refinement, generating an EmrAB-TolC reconstruction at a resolution of 3.14 Å. The resulting map quality for the EmrA and TolC sections was high, but the quality for the EmrB section was low. The AlphaFold2 55 -generated EmrB model was subsequently transformed into a map using e2pdb2mrc.py in EMAN2 56 and subsequently aligned to the EmrB map section within the EmrAB-TolC reconstruction. A focused mask on EmrB was applied to carry out 3D classification (without alignment), yielding five quality classes from 10. Aligning the 3D volumes and corresponding 19,180 particles from these classes using the Align 3D Maps program in CryoSPARC resulted in a homogeneous reconstruction of EmrAB-TolC at a resolution of 3.45 angstroms, with an improved map quality for the EmrB section. Local refinement was then carried out using a focused mask on the EmrAB portion, leading to a refined EmrAB-TolC reconstruction at 3.13 Å with improved map quality for the EmrAB sections (map-1 of pump-EA; and Extended Data Fig. 2c, d). Local refinement was also conducted using a focused mask on the TolC and α-helical hairpin domains of EmrA, leading to a refined reconstruction at 3.14 Å with better map quality for the EmrA-TolC sections (map-2 of pump-EA; and Extended Data Fig. 2c, d). A similar image processing procedure was employed for EmrAB-TolC pump-FA, which generated a reconstruction at 3.59 Å resolution (Extended Data Fig. 3c, d; Extended Data Table 2). Resolution estimation was accomplished via CryoSPARC using independently refined half-reconstructions. The criterion for Fourier shell correlation was set at 0.143. The ResMap wrapper in CryoSPARC was used to calculate variation in local resolution (Extended Data Fig. 2; Extended Data Fig. 3) 57 . Protein structure prediction and modelling AlphaFold3, AlphaFold2 and AF-Cluster were all employed for protein structure predictions. High-confidence predictions were selected based on their per-residue confidence score and corresponding predicted aligned error plot. Unless stated otherwise, the top rank of 5 predictions is shown and is visualized in ChimeraX Version 1.7. AlphaFold3 Beta was used to generate models of full-length EmrB (UniProt P0AEJ0) 36 . AlphaFold3-Multimer Beta was used to generate models of EmrAB, consisting of one molecule of full-length EmrB (UniProt P0AEJ0) with six molecules of full-length EmrA (UniProt P27303). AF-Cluster was used to predict multiple conformations of EmrB 37 . An MSA of EmrB (UniProt P0AEJ0) consisting of ~6000 sequences was first generated using ColabFold 58 . These sequences were then clustered into ~300 groups based on similarity using DBSCAN. A total of 16 clusters of sufficient size (at least 30 sequences) were then subsequently predicted using AlphaFold2. These conformations were visualized and analyzed using both ChimeraX and ChimeraUCSF. Model building and refinement The two half-maps of cryo-EM map-1 for EmrAB-TolC pump-EA were used to perform local map sharpening with DeepEMhancer 59 . The resulting full map was segmented into three TolC protomers, six EmrA protomers, and one EmrB protomer. A model of the β-barrel domain without β-CL, and the lipoyl domain, generated by AlphaFold3 36 , was fitted to the map of individual EmrA protomers using Chimera 60 . The α-helical hairpin domain, C-terminal α-helix, and β-CL of the β-barrel domain were manually built using Coot 61 . Each EmrA protomer model was subsequently refined with Rosetta 62 . A homology model of EmrB in the outward-open state was generated by AlphaFold3 36 . Chimera was used to fit the EBN and EBC domains of this homology model to the map. The EmrB model was subsequently refined using Rosetta. The two half-maps of cryo-EM map-2 for EmrAB-TolC pump-EA were used to conduct local map sharpening using DeepEMhancer. The structure of trimeric TolC, derived from the MacAB-TolC pump (PDB code: 5NIK), was docked into the TolC section of the modified map-2 using Chimera. To improve the local fit to the map, manual adjustments were made to the model using Coot. Models of individual components were fitted into the cryo-EM map-1 of EmrAB-TolC using Chimera. Adjustments were made to the entire EmrAB-TolC pump-EA model to reduce the clashes using ISOLDE 63 . The model was then undergone real-space refinement against this modified map in the Phenix package 64 . The cryo-EM map-1 of EmrAB-TolC pump-EA without global B-factor sharpening was then used to perform model-based local map sharpening with LocScale 65 . The two half-maps of cryo-EM map for EmrAB-TolC pump-FA were used to perform local map sharpening with DeepEMhancer. The refined model of EmrAB-TolC pump-EA was fitted into the cryo-EM map of pump-FA. The model section of the N-terminal TM helix of EmrA, generated by AlphaFold3, was fitted into the map of the individual EmrA protomer. Model sections without defined map were deleted. To improve the local fit to the map, manual adjustments were made to the model using Coot. The cryo-EM map of EmrAB-TolC pump-FA without global B-factor sharpening was then used to perform model-based local map sharpening with LocScale. Any Ramachandran outliers were manually corrected in Coot, and stereochemistry was ensured using MolProbity (Extended Data Table 2) 66 . Declarations Data availability Coordinates have been deposited in the Protein Data Bank (PDB) under PDB codes 8ZAL (EmrAB-TolC pump-EA) and 8ZAR (EmrAB-TolC pump-FA). Cryo-EM maps have been deposited in the Electron Microscopy Data Bank (EMD) under EMD codes EMD-39879 (EmrAB-TolC pump-EA) and EMD-39885 (EmrAB-TolC pump-FA). Acknowledgements: This work was supported by the National Key R&D Program of China (2022YFC2303200); National Natural Science Foundation of China (31971133 to D.D.; 32270064 and 92478118 to Y.C.), the Science and Technology Commission of Shanghai Municipality (19PJ1407900, 19JC1414000 and 22WZ2504100 to D.D.; 24ZR1493200 to Y.C.), and the Chinese Academy of Sciences (XDB0570000 and 176002GJHZ2022022MI to Y.C.). BFL was supported by ERC Advanced grant (742210) and a Wellcome Trust Investigator award (200873/Z/16/Z). Cryo-EM data were collected at the Bio-Electron Microscopy Facility of ShanghaiTech University with the assistance of Q. Sun, D. Liu, Z. Zhang, L. Wang and Y. Yang. We thank the Molecular Imaging Core Facility, the Molecular and Cell Biology Core Facility, and the Multi-Omics Core Facility at the School of Life Science and Technology for providing technical support. We are also grateful for the support of Lajos Kalmar of the MRC Toxicology Unit in the use of high-performance computing used in this study. Author contributions: Z.Z. performed cloning and overexpression of the EmrAB-TolC complex; Z.Z. and T.M. purified the EmrAB-TolC complex, prepared cryo-EM samples, collected cryo-EM data, determined structures, performed model building and refinement and prepared figures for the manuscript; X.G., W.S., H.Z., J.G., S.L., H.L. and Q.O. optimised in-column peptide-disc methods, prepared homemade graphene monolayer grids, and assisted the collection of cryo-EM data; R.D., H.J., Z.Z., T.M., X.G., S.Z. and W.S. performed antibiotics sensitivity assays; J.G., S.L., Z.Z. and T.M. performed drug-proton antiport assay; T.M. and H.L. carried out Western blot assay; Y.C. supervised antibiotics sensitivity assays and discussed project design; M.L.J. performed protein structure predictions and modelling;D.D. and B.L. conceived the project; D.D. designed and supervised all experiments; D.D. and B.L. wrote the manuscript. All the authors contributed to the data interpretation and manuscript preparation. Declaration of interests: The authors declare no competing interests. References Hodges, F. J., Torres, V. V. L., Cunningham, A. F., Henderson, I. R. & Icke, C. Redefining the bacterial Type I protein secretion system. Adv Microb Physiol 82 , 155-204 (2023). https://doi.org/10.1016/bs.ampbs.2022.10.003 Costa, T. R. et al. Secretion systems in Gram-negative bacteria: structural and mechanistic insights. Nat Rev Microbiol 13 , 343-359 (2015). https://doi.org/10.1038/nrmicro3456 Du, D., van Veen, H. W. & Luisi, B. F. Assembly and operation of bacterial tripartite multidrug efflux pumps. Trends Microbiol 23 , 311-319 (2015). https://doi.org/10.1016/j.tim.2015.01.010 Alav, I. et al. Structure, Assembly, and Function of Tripartite Efflux and Type 1 Secretion Systems in Gram-Negative Bacteria. Chem Rev 121 , 5479-5596 (2021). https://doi.org/10.1021/acs.chemrev.1c00055 Wang, Z. et al. An allosteric transport mechanism for the AcrAB-TolC multidrug efflux pump. Elife 6 (2017). https://doi.org/10.7554/eLife.24905 Du, D. et al. Structure of the AcrAB-TolC multidrug efflux pump. Nature 509 , 512-515 (2014). https://doi.org/10.1038/nature13205 Fitzpatrick, A. W. P. et al. Structure of the MacAB-TolC ABC-type tripartite multidrug efflux pump. Nat Microbiol 2 , 17070 (2017). https://doi.org/10.1038/nmicrobiol.2017.70 Du, D. et al. Multidrug efflux pumps: structure, function and regulation. Nat Rev Microbiol 16 , 523-539 (2018). https://doi.org/10.1038/s41579-018-0048-6 Yousefian, N. et al. Structural characterization of the EmrAB-TolC efflux complex from E. coli. Biochim Biophys Acta Biomembr 1863 , 183488 (2021). https://doi.org/10.1016/j.bbamem.2020.183488 Varela, M. F., Stephen, J., Bharti, D., Lekshmi, M. & Kumar, S. Inhibition of Multidrug Efflux Pumps Belonging to the Major Facilitator Superfamily in Bacterial Pathogens. Biomedicines 11 (2023). https://doi.org/10.3390/biomedicines11051448 Lin, M. F., Lin, Y. Y. & Lan, C. Y. Contribution of EmrAB efflux pumps to colistin resistance in Acinetobacter baumannii. J Microbiol 55 , 130-136 (2017). https://doi.org/10.1007/s12275-017-6408-5 Puertolas-Balint, F., Warsi, O., Linkevicius, M., Tang, P. C. & Andersson, D. I. Mutations that increase expression of the EmrAB-TolC efflux pump confer increased resistance to nitroxoline in Escherichia coli. J Antimicrob Chemother 75 , 300-308 (2020). https://doi.org/10.1093/jac/dkz434 Lomovskaya, O. L., K. Emr, an Escherichia coli locus for multidrug resistance. Cell Biology 89 (1992). Hinchliffe, P. et al. Structure of the periplasmic adaptor protein from a major facilitator superfamily (MFS) multidrug efflux pump. FEBS Lett 588 , 3147-3153 (2014). https://doi.org/10.1016/j.febslet.2014.06.055 Jiang, D. et al. Structure of the YajR transporter suggests a transport mechanism based on the conserved motif A. Proc Natl Acad Sci U S A 110 , 14664-14669 (2013). https://doi.org/10.1073/pnas.1308127110 Yin, Y., He, X., Szewczyk, P., Nguyen, T. & Chang, G. Structure of the Multidrug Transporter EmrD from Escherichia coli . Science 312 , 741-744 (2006). https://doi.org/10.1126/science.1125629 Nagarathinam, K. et al. Outward open conformation of a Major Facilitator Superfamily multidrug/H+ antiporter provides insights into switching mechanism. Nature Communications 9 (2018). https://doi.org/10.1038/s41467-018-06306-x Heng, J. et al. Substrate-bound structure of the E. coli multidrug resistance transporter MdfA. Cell Research 25 , 1060-1073 (2015). https://doi.org/10.1038/cr.2015.94 Majumder, P. et al. Dissection of Protonation Sites for Antibacterial Recognition and Transport in QacA, a Multi-Drug Efflux Transporter. J Mol Biol 431 , 2163-2179 (2019). https://doi.org/10.1016/j.jmb.2019.03.015 Remm, S. et al. Structural basis for triacylglyceride extraction from mycobacterial inner membrane by MFS transporter Rv1410. Nature Communications 14 (2023). https://doi.org/10.1038/s41467-023-42073-0 Tanabe, M. et al. The multidrug resistance efflux complex, EmrAB from Escherichia coli forms a dimer in vitro. Biochem Biophys Res Commun 380 , 338-342 (2009). https://doi.org/10.1016/j.bbrc.2009.01.081 Symmons, M. F., Marshall, R. L. & Bavro, V. N. Architecture and roles of periplasmic adaptor proteins in tripartite efflux assemblies. Frontiers in Microbiology 6 (2015). https://doi.org/10.3389/fmicb.2015.00513 Chun, E. et al. Fusion partner toolchest for the stabilization and crystallization of G protein-coupled receptors. Structure 20 , 967-976 (2012). https://doi.org/10.1016/j.str.2012.04.010 Gotzke, H. et al. The ALFA-tag is a highly versatile tool for nanobody-based bioscience applications. Nat Commun 10 , 4403 (2019). https://doi.org/10.1038/s41467-019-12301-7 Xie, P. et al. A fiducial-assisted strategy compatible with resolving small MFS transporter structures in multiple conformations using cryo-EM. Nat Commun 16 , 7 (2025). https://doi.org/10.1038/s41467-024-54986-5 Higgins, M. K., Bokma, E., Koronakis, E., Hughes, C. & Koronakis, V. Structure of the periplasmic component of a bacterial drug efflux pump. Proc Natl Acad Sci U S A 101 , 9994-9999 (2004). https://doi.org/10.1073/pnas.0400375101 Daury, L. et al. Tripartite assembly of RND multidrug efflux pumps. Nat Commun 7 , 10731 (2016). https://doi.org/10.1038/ncomms10731 Sulavik, M. C. et al. Antibiotic susceptibility profiles of Escherichia coli strains lacking multidrug efflux pump genes. Antimicrob Agents Chemother 45 , 1126-1136 (2001). https://doi.org/10.1128/AAC.45.4.1126-1136.2001 Su, C. C. et al. Crystal structure of the CusBA heavy-metal efflux complex of Escherichia coli. Nature 470 , 558-562 (2011). https://doi.org/10.1038/nature09743 Zhao, Y. et al. Crystal Structure of the E. coli Peptide Transporter YbgH. Structure 22 , 1152-1160 (2014). https://doi.org/10.1016/j.str.2014.06.008 Newstead, S. et al. Crystal structure of a prokaryotic homologue of the mammalian oligopeptide-proton symporters, PepT1 and PepT2. EMBO J 30 , 417-426 (2011). https://doi.org/10.1038/emboj.2010.309 Fowler, P. W. et al. Gating topology of the proton-coupled oligopeptide symporters. Structure 23 , 290-301 (2015). https://doi.org/10.1016/j.str.2014.12.012 Jin, J., Guffanti, A. A., Beck, C. & Krulwich, T. A. Twelve-Transmembrane-Segment (TMS) Version (ΔTMS VII-VIII) of the 14-TMS Tet(L) Antibiotic Resistance Protein Retains Monovalent Cation Transport Modes but Lacks Tetracycline Efflux Capacity. Journal of Bacteriology 183 , 2667-2671 (2001). https://doi.org/10.1128/jb.183.8.2667-2671.2001 Henderson, P. J. F. et al. Physiological Functions of Bacterial "Multidrug" Efflux Pumps. Chem Rev 121 , 5417-5478 (2021). https://doi.org/10.1021/acs.chemrev.0c01226 Drew, D., North, R. A., Nagarathinam, K. & Tanabe, M. Structures and General Transport Mechanisms by the Major Facilitator Superfamily (MFS). Chem Rev 121 , 5289-5335 (2021). https://doi.org/10.1021/acs.chemrev.0c00983 Abramson, J. et al. Accurate structure prediction of biomolecular interactions with AlphaFold 3. Nature (2024). https://doi.org/10.1038/s41586-024-07487-w Wayment-Steele, H. K. et al. Predicting multiple conformations via sequence clustering and AlphaFold2. Nature 625 , 832-839 (2024). https://doi.org/10.1038/s41586-023-06832-9 Sauve, S., Williamson, J., Polasa, A. & Moradi, M. Ins and Outs of Rocker Switch Mechanism in Major Facilitator Superfamily of Transporters. Membranes (Basel) 13 (2023). https://doi.org/10.3390/membranes13050462 Glavier, M. et al. Antibiotic export by MexB multidrug efflux transporter is allosterically controlled by a MexA-OprM chaperone-like complex. Nat Commun 11 , 4948 (2020). https://doi.org/10.1038/s41467-020-18770-5 Wu, H.-H., Symersky, J. & Lu, M. Structure and mechanism of a redesigned multidrug transporter from the Major Facilitator Superfamily. Scientific Reports 10 (2020). https://doi.org/10.1038/s41598-020-60332-8 Kim, J.-S. et al. Crystal Structure of a Soluble Fragment of the Membrane Fusion Protein HlyD in a Type I Secretion System of Gram-Negative Bacteria. Structure 24 , 477-485 (2016). https://doi.org/10.1016/j.str.2015.12.012 Teelucksingh, T. et al. A genetic platform to investigate the functions of bacterial drug efflux pumps. Nat Chem Biol 18 , 1399-1409 (2022). https://doi.org/10.1038/s41589-022-01119-y Angiulli, G. et al. New approach for membrane protein reconstitution into peptidiscs and basis for their adaptability to different proteins. Elife 9 (2020). https://doi.org/10.7554/eLife.53530 Carlson, M. L. et al. The Peptidisc, a simple method for stabilizing membrane proteins in detergent-free solution. Elife 7 (2018). https://doi.org/10.7554/eLife.34085 Fluman, N., Ryan, C. M., Whitelegge, J. P. & Bibi, E. Dissection of mechanistic principles of a secondary multidrug efflux protein. Mol Cell 47 , 777-787 (2012). https://doi.org/10.1016/j.molcel.2012.06.018 Bogdanov, M. Preparation of Uniformly Oriented Inverted Inner (Cytoplasmic) Membrane Vesicles from Gram-Negative Bacterial Cells. Methods Mol Biol 2715 , 159-180 (2024). https://doi.org/10.1007/978-1-0716-3445-5_10 Verkhovskaya, M. Preparation of Everted Membrane Vesicles from Escherichia coli Cells. Bio Protoc 7 , e2254 (2017). https://doi.org/10.21769/BioProtoc.2254 Schorb, M., Haberbosch, I., Hagen, W. J. H., Schwab, Y. & Mastronarde, D. N. Software tools for automated transmission electron microscopy. Nature Methods 16 , 471-477 (2019). https://doi.org/10.1038/s41592-019-0396-9 Li, X. et al. Electron counting and beam-induced motion correction enable near-atomic-resolution single-particle cryo-EM. Nature Methods 10 , 584-590 (2013). https://doi.org/10.1038/nmeth.2472 Rohou, A. & Grigorieff, N. CTFFIND4: Fast and accurate defocus estimation from electron micrographs. Journal of Structural Biology 192 , 216-221 (2015). https://doi.org/10.1016/j.jsb.2015.08.008 Scheres, S. H. RELION: implementation of a Bayesian approach to cryo-EM structure determination. J Struct Biol 180 , 519-530 (2012). https://doi.org/10.1016/j.jsb.2012.09.006 Zivanov, J. et al. New tools for automated high-resolution cryo-EM structure determination in RELION-3. Elife 7 (2018). https://doi.org/10.7554/eLife.42166 Punjani, A., Rubinstein, J. L., Fleet, D. J. & Brubaker, M. A. cryoSPARC: algorithms for rapid unsupervised cryo-EM structure determination. Nat Methods 14 , 290-296 (2017). https://doi.org/10.1038/nmeth.4169 Wright, N. J. et al. Methotrexate recognition by the human reduced folate carrier SLC19A1. Nature 609 , 1056-1062 (2022). https://doi.org/10.1038/s41586-022-05168-0 Jumper, J. et al. Highly accurate protein structure prediction with AlphaFold. Nature 596 , 583-589 (2021). https://doi.org/10.1038/s41586-021-03819-2 Tang, G. et al. EMAN2: an extensible image processing suite for electron microscopy. J Struct Biol 157 , 38-46 (2007). https://doi.org/10.1016/j.jsb.2006.05.009 Kucukelbir, A., Sigworth, F. J. & Tagare, H. D. Quantifying the local resolution of cryo-EM density maps. Nature Methods 11 , 63-65 (2014). https://doi.org/10.1038/nmeth.2727 Mirdita, M. et al. ColabFold: making protein folding accessible to all. Nature Methods 19 , 679-682 (2022). https://doi.org/10.1038/s41592-022-01488-1 Sanchez-Garcia, R. et al. DeepEMhancer: a deep learning solution for cryo-EM volume post-processing. Commun Biol 4 , 874 (2021). https://doi.org/10.1038/s42003-021-02399-1 Pettersen, E. F. et al. UCSF Chimera--a visualization system for exploratory research and analysis. J Comput Chem 25 , 1605-1612 (2004). https://doi.org/10.1002/jcc.20084 Emsley, P., Lohkamp, B., Scott, W. G. & Cowtan, K. Features and development of Coot. Acta Crystallogr D Biol Crystallogr 66 , 486-501 (2010). https://doi.org/10.1107/S0907444910007493 Wang, R. Y. et al. Automated structure refinement of macromolecular assemblies from cryo-EM maps using Rosetta. Elife 5 (2016). https://doi.org/10.7554/eLife.17219 Croll, T. I. ISOLDE: a physically realistic environment for model building into low-resolution electron-density maps. Acta Crystallogr D Struct Biol 74 , 519-530 (2018). https://doi.org/10.1107/S2059798318002425 Liebschner, D. et al. Macromolecular structure determination using X-rays, neutrons and electrons: recent developments in Phenix. Acta Crystallogr D Struct Biol 75 , 861-877 (2019). https://doi.org/10.1107/S2059798319011471 Jakobi, A. J., Wilmanns, M. & Sachse, C. Model-based local density sharpening of cryo-EM maps. Elife 6 (2017). https://doi.org/10.7554/eLife.27131 Williams, C. J. et al. MolProbity: More and better reference data for improved all-atom structure validation. Protein Sci 27 , 293-315 (2018). https://doi.org/10.1002/pro.3330 Additional Declarations There is NO Competing Interest. Supplementary Files ExtendedData.docx Cite Share Download PDF Status: Published Journal Publication published 16 Mar, 2026 Read the published version in Nature Communications → Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-5901545","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":414747919,"identity":"bf7804d1-fd7f-43c7-a0c6-bb277b95a2a1","order_by":0,"name":"Ben Luisi","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAAuklEQVRIiWNgGAWjYBACA+bDjQ8SKuD8BCK0sCU2G3w4Q6KWNsmZbaRoMWdjbJDmnWcnr9vA/PADY1saYS2WbYwNxrzbkg23HWAzlmBsyyHCYfcbG5J5tx1g3HaAwYyBsa2CCC3HGBsO8845YL/tAPs3orU0Ns5sOJC47QAPyBYiHAb0SzPDh2PJydsO8xRLJJwjwvvmbMzHfyTU2NluO96+8cOHsmTCWhCAmYGoWBkFo2AUjIJRQAwAANgKOdx+URftAAAAAElFTkSuQmCC","orcid":"https://orcid.org/0000-0003-1144-9877","institution":"University of Cambridge","correspondingAuthor":true,"prefix":"","firstName":"Ben","middleName":"","lastName":"Luisi","suffix":""},{"id":414747920,"identity":"eedb9878-f8d5-41c6-896b-241b2b8087d3","order_by":1,"name":"Dijun Du","email":"","orcid":"https://orcid.org/0009-0008-8825-4932","institution":"ShanghaiTech University","correspondingAuthor":false,"prefix":"","firstName":"Dijun","middleName":"","lastName":"Du","suffix":""},{"id":414747921,"identity":"62370b28-fc44-4b76-a211-b17aed4a37c8","order_by":2,"name":"Yanjie Chao","email":"","orcid":"https://orcid.org/0000-0002-5735-7954","institution":"Chinese Academy of Sciences","correspondingAuthor":false,"prefix":"","firstName":"Yanjie","middleName":"","lastName":"Chao","suffix":""},{"id":414747922,"identity":"52ef909c-4853-4383-a3a2-deba965d00df","order_by":3,"name":"Zhaojun Zhong","email":"","orcid":"","institution":"ShanghaiTech University","correspondingAuthor":false,"prefix":"","firstName":"Zhaojun","middleName":"","lastName":"Zhong","suffix":""},{"id":414747923,"identity":"3353e830-ee40-496a-861c-9b6fec964393","order_by":4,"name":"Tuerxunjiang Maimaiti","email":"","orcid":"https://orcid.org/0009-0007-0309-9682","institution":"ShanghaiTech University","correspondingAuthor":false,"prefix":"","firstName":"Tuerxunjiang","middleName":"","lastName":"Maimaiti","suffix":""},{"id":414747924,"identity":"6d88aad0-c48f-4805-87d3-9c106de0e589","order_by":5,"name":"Matthew Jackson","email":"","orcid":"","institution":"University of Cambridge","correspondingAuthor":false,"prefix":"","firstName":"Matthew","middleName":"","lastName":"Jackson","suffix":""},{"id":414747925,"identity":"d6250ac5-4828-44a3-8cbc-5acda75ce447","order_by":6,"name":"Rui Dong","email":"","orcid":"","institution":"Shanghai Institute of Immunity and Infection, CAS","correspondingAuthor":false,"prefix":"","firstName":"Rui","middleName":"","lastName":"Dong","suffix":""},{"id":414747926,"identity":"ba3472c4-3e3d-4725-b70a-ef11222215c2","order_by":7,"name":"Xueyan Gao","email":"","orcid":"","institution":"ShanghaiTech University","correspondingAuthor":false,"prefix":"","firstName":"Xueyan","middleName":"","lastName":"Gao","suffix":""},{"id":414747927,"identity":"c2ad3092-e998-45f8-8739-dfcf36d2ceab","order_by":8,"name":"Jinliang Guo","email":"","orcid":"https://orcid.org/0009-0001-3375-0196","institution":"ShanghaiTech University","correspondingAuthor":false,"prefix":"","firstName":"Jinliang","middleName":"","lastName":"Guo","suffix":""},{"id":414747928,"identity":"acb73ddc-7810-414f-ae1d-4435616b6fb8","order_by":9,"name":"Shangrong Li","email":"","orcid":"https://orcid.org/0009-0004-5390-1504","institution":"ShanghaiTech University","correspondingAuthor":false,"prefix":"","firstName":"Shangrong","middleName":"","lastName":"Li","suffix":""},{"id":414747929,"identity":"1f6cb78b-ed6a-45cc-a9b8-3ed0580f02e9","order_by":10,"name":"Wenyu Shang","email":"","orcid":"","institution":"ShanghaiTech University","correspondingAuthor":false,"prefix":"","firstName":"Wenyu","middleName":"","lastName":"Shang","suffix":""},{"id":414747930,"identity":"427404d0-b106-4615-bc71-3031326a1ef1","order_by":11,"name":"Huajun Liu","email":"","orcid":"","institution":"ShanghaiTech University","correspondingAuthor":false,"prefix":"","firstName":"Huajun","middleName":"","lastName":"Liu","suffix":""},{"id":414747931,"identity":"b0378225-cb34-4757-a760-7ec4fc0cf75a","order_by":12,"name":"Hongnian Jiang","email":"","orcid":"","institution":"Chinese Academy of Sciences","correspondingAuthor":false,"prefix":"","firstName":"Hongnian","middleName":"","lastName":"Jiang","suffix":""},{"id":414747932,"identity":"43adb373-ac5a-454d-b5ae-59dcbbd650c4","order_by":13,"name":"Shuo Zhang","email":"","orcid":"","institution":"Chinese Academy of Sciences","correspondingAuthor":false,"prefix":"","firstName":"Shuo","middleName":"","lastName":"Zhang","suffix":""},{"id":414747933,"identity":"efc9fabb-be22-43b4-8b90-d18bc9828d11","order_by":14,"name":"Huimin Zhao","email":"","orcid":"","institution":"ShanghaiTech University","correspondingAuthor":false,"prefix":"","firstName":"Huimin","middleName":"","lastName":"Zhao","suffix":""},{"id":414747934,"identity":"3fbdd2e3-1f25-4ca0-821c-508c24e4b45e","order_by":15,"name":"Qing Ouyang","email":"","orcid":"","institution":"Shanghai Tech","correspondingAuthor":false,"prefix":"","firstName":"Qing","middleName":"","lastName":"Ouyang","suffix":""}],"badges":[],"createdAt":"2025-01-25 12:10:10","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-5901545/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-5901545/v1","draftVersion":[],"editorialEvents":[{"content":"https://doi.org/10.1038/s41467-026-70500-5","type":"published","date":"2026-03-16T04:00:00+00:00"}],"editorialNote":"","failedWorkflow":false,"files":[{"id":76218255,"identity":"7fa06fe2-c4ef-4a94-bd8c-097a3f38e640","added_by":"auto","created_at":"2025-02-13 14:49:54","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":827026,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eOverall structure of the EmrAB-TolC efflux pump. a, \u003c/strong\u003eCryoEM\u003cstrong\u003e \u003c/strong\u003emap of pump-EA (Map-1). Subunits are colour-coded as in \u003cstrong\u003ec\u003c/strong\u003e.\u003cstrong\u003e b,\u003c/strong\u003eRibbon representation of the pump with the color code shown in \u003cstrong\u003ec\u003c/strong\u003e.\u003cstrong\u003e \u003c/strong\u003eThe TolC trimer spans the outer membrane, with its helical end extending into the periplasm, where it engages the EmrA hexamer. The β-barrel domains of EmrA interact with the periplasmic portion of EmrB.\u003c/p\u003e","description":"","filename":"1.png","url":"https://assets-eu.researchsquare.com/files/rs-5901545/v1/c774cd8fec77b6d3feb730a0.png"},{"id":76218076,"identity":"79f6c195-af0d-46bd-837b-ed667e380979","added_by":"auto","created_at":"2025-02-13 14:41:52","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":567186,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eStructure of EmrA. a,\u003c/strong\u003e Linear representation of EmrA (the colour code in \u003cstrong\u003ea\u003c/strong\u003e is also used for the domains and subdomains in \u003cstrong\u003eb\u003c/strong\u003e). \u003cstrong\u003eb,\u003c/strong\u003e Ribbon representation of EmrA protomer 4. \u003cstrong\u003ec,\u003c/strong\u003e Cryo-EM map of the EmrA hexamer with the fitted model (left). Protomers are colour-coded accordingly (right). The map of the N-terminal TM helix of EmrA protomer 4 spans the entire IM, while the maps of the other five N-terminal TM helices are approximately two-thirds shorter. The speculative model portions of these five N-terminal TM helices without defined map are coloured light gray.\u003c/p\u003e","description":"","filename":"2.png","url":"https://assets-eu.researchsquare.com/files/rs-5901545/v1/1c5209582680ea4650a8b10d.png"},{"id":76218069,"identity":"44bcdc03-6a88-42ef-a2d3-9dd768d01afe","added_by":"auto","created_at":"2025-02-13 14:41:51","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":438623,"visible":true,"origin":"","legend":"\u003cp\u003eStructure and function of EmrB. a, Linear representation of EmrB (the colour code in A is also used for the domains and subdomains in b). b, c, Ribbon representations of EmrB viewed parallel to the membrane and perpendicular to the membrane on the periplasmic side, respectively. d, Organisation of critical residues potentially involved in H\u003csup\u003e+\u003c/sup\u003e translocation and coupling. e, Mutants of EmrA/EmrB exhibit loss of resistance to nalidixic acid, nitroxoline and CCCP. Survival assays of \u003cem\u003eE. coli\u003c/em\u003e C43(DE3)\u003cem\u003e ΔacrAB\u003c/em\u003e expressing wild-type EmrAB-TolC, EmrA-TolC, EmrB-TolC, EmrAB\u003csub\u003eD29N\u003c/sub\u003e-TolC, or EmrAB\u003csub\u003eR109A\u003c/sub\u003e-TolC, were carried out as in Extended Data Fig. 1a. \u0026nbsp;\u003c/p\u003e","description":"","filename":"3.png","url":"https://assets-eu.researchsquare.com/files/rs-5901545/v1/e1dcbce869d855cac8e26164.png"},{"id":76218063,"identity":"a8eb0be0-614d-4cc0-be3e-50f0a5290275","added_by":"auto","created_at":"2025-02-13 14:41:51","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":1563131,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eInteractions between EmrA and EmrB. a, \u003c/strong\u003eMap and model showing the interfacial contacts between EmrA and EmrB viewed parallel to the membrane. Protomers 3 and 4 of EmrA are omitted for clarity. \u003cstrong\u003eb,\u003c/strong\u003e Surface representations of the β-barrel domains of EmrA and the AssB subdomain of EmrB, showing contacts between β-CLs and AssB, as well as CavEA and β-CL1PH (viewed perpendicular to the membrane at the inner membrane distal (left) and proximal (middle) sides, respectively). Domains and subdomains are colour-coded accordingly (right). \u003cstrong\u003ec,\u003c/strong\u003e The transport pathway in the EmrB and EmrA β-barrel domains, viewed perpendicular (left) and parallel (right) to the membrane. The cavities CavEB and CavEA and the path β-CL1PH are coloured red and were calculated using Caver 3.0 \u003csup\u003e43\u003c/sup\u003e. \u003cstrong\u003ed,\u003c/strong\u003e Mutants of EmrA exhibit loss of resistance to nalidixic acid, nitroxoline, and CCCP. Survival assays of \u003cem\u003eE. coli\u003c/em\u003e C43(DE3)\u003cem\u003e ΔacrAB \u003c/em\u003eexpressing wild-type EmrAB-TolC, EmrA\u003csub\u003eL302Q/303Q\u003c/sub\u003e-EmrB-TolC, or EmrA\u003csub\u003eI313Q/V315T/V316T\u003c/sub\u003e-EmrB-TolC, were carried out as in Extended Data Fig. 1a. “WT” and “EV” denote wild-type EmrAB-TolC and the empty vector containing the negative control, respectively. \u003cstrong\u003ee\u003c/strong\u003e, Western blot assays demonstrate that the expression levels of EmrA mutants in the EmrAB-TolC pumps are similar to those of the wild type ones, and confirm that these pump components are properly localized in the membrane.\u003c/p\u003e","description":"","filename":"4.png","url":"https://assets-eu.researchsquare.com/files/rs-5901545/v1/8b4c2380fc7cfe6cf2977614.png"},{"id":76218090,"identity":"a040080e-48d9-4c16-b8c4-970a431dc9ad","added_by":"auto","created_at":"2025-02-13 14:41:52","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":169327,"visible":true,"origin":"","legend":"\u003cp\u003eConformational changes of EmrAB. a, Alignment of the experimentally determined structure of EmrB with the top-rank prediction of EmrB using AlphaFold3. b, AlphaFold3 and AF-Cluster predict EmrB in its inward-open conformation. This conformation has a large cytosolic binding cleft, CavEB\u003csub\u003ein\u003c/sub\u003e. c, Snapshots of the hydrophobic inner section of EmrAB highlight differential cavities involved in the transport of antibiotics. d, AlphaFold2 structures with high confidence scores were predicted from discrete sequence clusters and overlaid using ChimeraX. These structures reveal the conformational dynamics of EmrB as it transitions between an outward-open to inward-open state through changes in the positions of helices H1-H6 about a central pivot point. e, D29 exhibits 3 polar contacts in its outward-open state, in comparison to just 2 contacts in the inward-open state. f, Mutants of EmrB exhibit loss of resistance to nalidixic acid, nitroxoline and CCCP. Survival assays of \u003cem\u003eE. coli\u003c/em\u003e C43(DE3)\u003cem\u003e ΔacrAB\u003c/em\u003e expressing wild-type EmrAB-TolC, EmrAB\u003csub\u003eN62A\u003c/sub\u003e-TolC, EmrAB\u003csub\u003eV147A\u003c/sub\u003e-TolC, EmrAB\u003csub\u003eV148A\u003c/sub\u003e-TolC, or EmrAB\u003csub\u003eI288A\u003c/sub\u003e-TolC, were carried out as in Extended Data Fig. 1a.\u0026nbsp; g, Western blot assays demonstrate that the expression levels of EmrB mutants in the EmrAB-TolC pump are similar to those of the wild type and confirm that these pump components are properly localized in the membrane. h, Proton transport by everted membrane vesicles expressing wild-type EmrAB-TolC or EmrAB\u003csub\u003eD29N\u003c/sub\u003e-TolC/EmrAB\u003csub\u003eV147A\u003c/sub\u003e-TolC mutants, respectively. The transmembrane ΔpH was followed by measuring 9-Amino-6-Chloro-2-Methoxyacridine (ACMA) fluorescence. Lactate and NH\u003csub\u003e4\u003c/sub\u003eCl were used to generate or dissipate ΔpH, respectively. Nalidixic acid was added after 1 min. “WT” and “EV” denote wild-type EmrAB-TolC and the empty vector containing the negative control, respectively.\u003c/p\u003e","description":"","filename":"5.png","url":"https://assets-eu.researchsquare.com/files/rs-5901545/v1/3754474458cdb42c18476ac5.png"},{"id":76218098,"identity":"3e224cd3-05e1-4fba-a1fa-8a55f7cc3699","added_by":"auto","created_at":"2025-02-13 14:41:53","extension":"png","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":212969,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eSchematic cartoon diagrams of the one-step drug transport mechanism of EmrAB-TolC.\u0026nbsp; a, \u003c/strong\u003eThe resting apo tripartite pump with EmrB in the outward-open state. Helix H\u003csub\u003eY\u003c/sub\u003e in the AssB subdomain of EmrB anchors onto a hydrophobic ring-1 in the AssA subdomain of EmrA to facilitate pump assembly. Engagement between TolC and EmrA directly opens the periplasmic aperture of TolC. Exposure to a low pH periplasmic space leads to protonation of the acidic residue D29 of EmrB. \u003cstrong\u003eb,\u003c/strong\u003e The protonation of D29 changes its salt bridge and hydrogen bond network with residues R109, Q112, and N174, transforms the conformation of EmrB to an inward-open state, allowing H\u003csup\u003e+\u003c/sup\u003e release from D29 to the cytosol, and drug binding to a recognition site. \u003cstrong\u003ec,\u003c/strong\u003e Drug binding to EmrB triggers a structural transition that exposes the binding pocket to the periplasm. The drug is then released into CavEA\u003csub\u003eout\u003c/sub\u003e. \u003cstrong\u003ed,\u003c/strong\u003e The protonation of D29 returns EmrB to an inward-open state again. As a result, the periplasmic regions between the transmembrane helices of EmrB shift toward the central axis, partially filling the space in CavEA. The contraction of CavEA\u003csub\u003eout\u003c/sub\u003e propels substrates along the β-CL1PH-CavLP pathway toward the nanochannel formed by EmrA and TolC.\u003c/p\u003e","description":"","filename":"6.png","url":"https://assets-eu.researchsquare.com/files/rs-5901545/v1/67592be753ec77a775d2bc64.png"},{"id":108476810,"identity":"b2c6b1de-37ca-4a9e-8cbf-ff5abb7e1119","added_by":"auto","created_at":"2026-05-05 07:05:42","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":4760113,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-5901545/v1/58ba6ed5-949d-4207-b33c-426441917c37.pdf"},{"id":76218245,"identity":"7b84f0e1-cbe9-49a8-9221-1ecedd3b98a3","added_by":"auto","created_at":"2025-02-13 14:49:52","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":18077324,"visible":true,"origin":"","legend":"","description":"","filename":"ExtendedData.docx","url":"https://assets-eu.researchsquare.com/files/rs-5901545/v1/a5158944271277587d872f9d.docx"}],"financialInterests":"There is \u003cb\u003eNO\u003c/b\u003e Competing Interest.","formattedTitle":"One-step drug transport across two membranes of Gram-negative bacteria by an MFS tripartite assembly","fulltext":[{"header":"Introduction","content":"\u003cp\u003eAntimicrobial resistance of Gram-negative bacteria is one of the most devastating issues in modern healthcare systems worldwide. The complex architecture of the cell envelope in Gram-negative bacteria presents a formidable obstacle for the penetration and transport of antimicrobial drugs. As one way to develop drug resistance, these organisms have evolved an array of specialized macromolecular nanomachines capable of extruding internalized drugs back to the environments \u003csup\u003e\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e,\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e\u003c/sup\u003e, for example the tripartite efflux pumps, double membrane-spanning nanomachines that are closely related with type I secretion systems (T1SSs) \u003csup\u003e\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e,\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e\u003c/sup\u003e. To cross the double membranes of Gram-negative bacteria, substrates may be exported directly from the cytosol to the outer membrane (OM) in a one-step manner, or indirectly in a two-step fashion in which substrates are first moved to the periplasm through a transporter system on the inner membrane (IM), and then translocated across the OM by another \u003csup\u003e\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e\u003c/sup\u003e. Of three classes of tripartite efflux pumps, two operate on the two-step transport mechanism \u003csup\u003e\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e\u003c/sup\u003e, including the resistance-nodulation-cell division (RND) types exemplified by AcrAB-TolC \u003csup\u003e\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e,\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e\u003c/sup\u003e and the ATP-binding cassette (ABC) types represented by MacAB-TolC \u003csup\u003e\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e\u003c/sup\u003e. The MFS-type tripartite efflux pumps are believed to employ a one-step transport mechanism like T1SSs, however the molecular details of the one-step process remain unclear \u003csup\u003e\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e,\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eThe MFS-type tripartite efflux pumps are conserved in many Gram-negative bacterial pathogens \u003csup\u003e\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e,\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e\u003c/sup\u003e. The prototype MFS-type EmrAB-TolC pump of \u003cem\u003eEscherichia coli\u003c/em\u003e plays an essential role in mediating multi-drug resistance and contributes to virulence phenotypes \u003csup\u003e\u003cspan additionalcitationids=\"CR10 CR11 CR12\" citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e\u003c/sup\u003e. This pump consists of three components: a TolC OM channel, an EmrB IM transporter, and an EmrA periplasmic adaptor which bridges the two transmembrane proteins \u003csup\u003e\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e\u003c/sup\u003e. Understanding of the overall structure and organization of the EmrAB-TolC pump is currently limited to the structural data of individual, separate components of its homologues \u003csup\u003e\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e,\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e\u003c/sup\u003e. The \u003cem\u003eAquifex aeolicus\u003c/em\u003e EmrA structure exhibits a linear arrangement of α-helical hairpin, lipoyl, and β-barrel domains. A conserved loop in the β-barrel domain is much longer than that of other adaptors, which is disordered in the determined structure \u003csup\u003e\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e\u003c/sup\u003e. EmrB belongs to the DHA2 transporter family and, unlike the well-characterized DHA1 family for which structures in various conformations are available \u003csup\u003e\u003cspan additionalcitationids=\"CR16 CR17\" citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e\u003c/sup\u003e, the DHA2 family has been less well characterized. However, recent studies on an analogous DHA2 transporter QacA have revealed critical acidic residues essential for substrate recognition and transport \u003csup\u003e\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e\u003c/sup\u003e. Structural analysis of another DHA2 transporter MHAS2168 from \u003cem\u003eMycobacterium hassiacum\u003c/em\u003e revealed the extension of two transmembrane (TM) helices, TM 11 and TM12, into the periplasm, which might interact with a lipoprotein for substrate translocation \u003csup\u003e\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e\u003c/sup\u003e. The stoichiometry of the EmrAB-TolC assembly remains undefined, with studies reporting conflicting models of either 3:6:1 or 3:6:2 TolC:EmrA:EmrB ratios \u003csup\u003e\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e,\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e,\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e\u003c/sup\u003e. The high-resolution structure of the full pump assembly and the mechanism of the pump are currently unknown.\u003c/p\u003e \u003cp\u003eIn this study, we present the cryo-EM structure of the \u003cem\u003eE. coli\u003c/em\u003e EmrAB-TolC assembly for the first-time, revealing how the pump confers multi-drug resistance to antibiotics. The pump structure pinpoints several key residues that are required for drug export, mutation of which abrogate bacterial resistance to multiple antibiotics. The architecture of the EmrAB-TolC pump uncovers a one-step mechanism that directly transport antimicrobial drugs across the entire cell envelope of Gram-negative bacteria, offering new insights on the mechanisms of antimicrobial resistance.\u003c/p\u003e"},{"header":"Results","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003eEngineering stable and functional EmrAB-TolC complexes\u003c/h2\u003e \u003cp\u003eYousefian \u003cem\u003eet al.\u003c/em\u003e demonstrated that the native tripartite EmrAB-TolC complex can be obtained, but the yield is insufficient for cryo-EM analysis \u003csup\u003e\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e\u003c/sup\u003e, because the protein complex is highly susceptible to both dissociation and precipitation. To stabilize the pump assembly, we first modified EmrB by fusing a thermostabilized apocytochrome b562RIL (BRIL) protein and an ALFA tag to the N- and C-termini of the protein \u003csup\u003e\u003cspan additionalcitationids=\"CR24\" citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e\u003c/sup\u003e, respectively. We then engineered an EmrB-EmrA fusion protein with a flexible linker that connected the modified EmrB and EmrA. We observed that the N-terminus of EmrA and the C-terminus of EmrB are likely adjacent to the cytoplasmic side, and EmrA possesses a single N-terminal TM helix. Consequently, EmrA was fused to the C-terminus of modified EmrB via a flexible polyglycine-serine linker, thereby maintaining the correct membrane topology of the pump components.\u003c/p\u003e \u003cp\u003eCo-expression of the fused EmrB-EmrA protein, free EmrA, and TolC led to the formation of a complex (engineered pump-FA). However, despite successful expression, the complex tended to dissociate or precipitate during purification, resulting in protein concentrations of only\u0026thinsp;~\u0026thinsp;0.1 mg/mL. Therefore, the N-terminal segment (residues 1\u0026minus;47) of free EmrA (containing the N-terminal TM helix) was replaced by the AcrA signal sequence (residues 1\u0026minus;27) containing residue C25, which can be modified by a palmitoyl acid chain to anchor EmrA to the IM \u003csup\u003e\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e,\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e\u003c/sup\u003e. The modified free EmrA, EmrB-EmrA fusion protein and TolC were successfully expressed and copurified, and the complex (engineered pump-EA) exhibited modest stability during purification, achieving a concentration of ~\u0026thinsp;0.7 mg/ml.\u003c/p\u003e \u003cp\u003eThe genomic \u003cem\u003eemrAB\u003c/em\u003e is expressed poorly in \u003cem\u003eE. coli\u003c/em\u003e under laboratory growth conditions, and its knockout does not significantly affect resistance \u003csup\u003e\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e\u003c/sup\u003e. However, mutations that increase the expression of the pump lead to increased resistance \u003csup\u003e\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e\u003c/sup\u003e. Therefore, we induced the expression of the pump to carry out survival assays, which revealed that the wild-type EmrAB-TolC pump conferred drug resistance to nitroxoline, nalidixic acid and CCCP (Extended Data Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003ea). Induced expression of the engineered EmrAB-TolC pump-EA pump showed that it conferred reduced resistance to nitroxoline, indicating that the complex was functional \u003cem\u003ein vivo\u003c/em\u003e. The EmrAB-TolC pump-FA, however, exhibits no resistance activity (Extended Data Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eb).\u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003eQuaternary structure of the EmrAB-TolC pump\u003c/h3\u003e\n\u003cp\u003eThe structure of EmrAB-TolC was solved using single-particle cryo-EM. Three maps were reconstructed: two maps for pump-EA achieved overall resolutions of 3.13 \u0026Aring; and 3.14 \u0026Aring;, respectively; and one map for pump-FA reached 3.59 \u0026Aring; resolution (Extended Data Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003ea-c; Extended Data Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003ea-c). Local resolution examination of these maps indicated a resolution ranging between 3.0 \u0026Aring; to 7.0 \u0026Aring; (Extended Data Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003ed; Extended Data Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003ed). The map quality of EmrAB-TolC pump-FA was insufficient for reliably constructing the full pump structure, whereas the maps for pump-EA were of sufficient quality to allow building models for residues 47\u0026minus;390 of EmrA and residues 13\u0026minus;498 of EmrB (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e; Extended Data Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003ee). Nevertheless, the model derived from the EmrAB-TolC pump-EA map fits well within the EmrAB-TolC pump-FA map. The models built from both maps show minimal differences, indicating that the pump-EA variant does not disrupt the assembly or structure of the complex. The map for pump-FA was used to construct models for the extra residues 17\u0026minus;46 of EmrA (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003ea-c; Extended Data Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eThe pump has an elongated shape consisting of a TolC trimer, an EmrA hexamer and an EmrB monomer (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e; Extended Data Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e). In the case of pump-FA, the model of the N-terminal TM helix of the EmrA protomer 4 spans the entire IM, while the partial models of the other five EmrA protomers are shorter (Extended Data Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e; Extended Data Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eThe positions of EmrB and TolC delineate the IM and OM boundaries, respectively, with the long axis of the pump assembly extending\u0026thinsp;~\u0026thinsp;320 \u0026Aring; into the periplasm (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e; Extended Data Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e). This is comparable to the dimensions observed for other types of tripartite assemblies, such as AcrAB-TolC and MacAB-TolC\u003csup\u003e\u003cspan additionalcitationids=\"CR6\" citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e\n\u003ch3\u003eThe EmrA hexamer forms a nanochannel with an AssA subdomain for pump assembly\u003c/h3\u003e\n\u003cp\u003eThe cryo-EM map revealed five sequentially ordered structural units in EmrA: the singular N-terminal TM helix and the α-helical hairpin, lipoyl, and β-barrel domains, as well as a C-terminal α-helix (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003ea-c; Extended Data Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003e). The N-terminal TM helix anchors the four periplasmic units to the IM (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eb, c). Adaptor proteins, such as AcrA and MacA, utilize a MP domain for interaction with the periplasmic extension of IM transporters\u003csup\u003e\u003cspan additionalcitationids=\"CR6\" citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e\u003c/sup\u003e. EmrA and its homologues, as well as the adaptor proteins of T1SS, however, lack such a MP domain (Extended Data Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003e; Extended Data Fig.\u0026nbsp;7). The α-helical hairpin domain structure resembles that of \u003cem\u003eA. aeolicus\u003c/em\u003e EmrA, although it is one-third shorter (Extended Data Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003ea, b) \u003csup\u003e\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e\u003c/sup\u003e. The structures of both the lipoyl and β-barrel domains are similar to those of other adaptors (Extended Data Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003e; Extended Data Fig.\u0026nbsp;7). A notable difference is the presence of a unique loop (termed β-CL), comprising residues 294\u0026minus;317, in the β-barrel domain of EmrA (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eb; Extended Data Fig.\u0026nbsp;8a), which is conserved among EmrA homologues (Extended Data Fig.\u0026nbsp;9a). A long coil follows the C-terminus of the β-barrel domain, folding back to engage with the side of the adjacent lipoyl domain in the pump assembly. This coil precedes an α-helix at the terminus (residues 375\u0026minus;390) that, together with the lower portion of the α-helical hairpin, forms a three-stranded coiled-coil, mimicking the α-helical hairpin domain of CusB \u003csup\u003e\u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eThe α-helical hairpin, lipoyl and β-barrel domains of EmrA assemble into a hexameric structure reminiscent of AcrA and MacA within the MacAB-TolC and AcrAB-TolC pumps (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003ec; Extended Data Fig.\u0026nbsp;7) \u003csup\u003e\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e,\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e\u003c/sup\u003e. The helical hairpin domains create an α-helical barrel nanochannel, referred to as NanoHP. The lipoyl domains form a ring encompassing an internal chamber that extends the nanochannel (named CavLP). The analogous hexameric assembly of MacA presents a gating ring formed by conserved glutamine residues in the lipoyl domain\u003csup\u003e\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e\u003c/sup\u003e, but this is notably absent in the EmrA assembly. The six β-CLs within the β-barrel domains (β-CL1 to β-CL6) vary in conformation, creating a structural unit which we refer to as assembly subdomain A (AssA), which forms the wall of a chamber (named CavEA) (Extended Data Fig.\u0026nbsp;8a-c). Residues L302 and L303 in the β-CL hexamer form a hydrophobic ring-1 on the IM distal side of CavEA, providing a narrow entryway to CavLP. Conversely, residues I313, V315 and V316 generate a hydrophobic ring, ring 2, on the IM proximal side of CavEA, potentially situated on the surface of the cellular membrane (Extended Data Fig.\u0026nbsp;8b, c; Extended Data Fig.\u0026nbsp;9a). The β-CL1 contains an alternative route (termed β-CL1PH) linking the chambers CavEA and CavLP (Extended Data Fig.\u0026nbsp;8c).\u003c/p\u003e\n\u003ch3\u003eA monomer of EmrB containing an AssB subdomain mediates pump assembly\u003c/h3\u003e\n\u003cp\u003eEmrB, which belongs to the DHA2 family, comprises 14 transmembrane helices, of which H1\u0026minus;H12 adopt the characteristic fold of MFS transporters such as MdfA and EmrD \u003csup\u003e\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e,\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e\u003c/sup\u003e. The six-helix bundles, formed by H1\u0026minus;H6 (termed EBN) and H7\u0026minus;H12 (termed EBC), create a V-shaped transporter (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003ea-c) with a pseudo-twofold symmetry axis perpendicular to the membrane plane. Two more TM helices, H\u003csub\u003eA\u003c/sub\u003e and H\u003csub\u003eB\u003c/sub\u003e, are inserted into the cytoplasmic loop bridging the EBN and EBC domains, but against the periphery of the protein core, similar to YbgH and PepTso \u003csup\u003e\u003cspan additionalcitationids=\"CR31\" citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e\u003c/sup\u003e. The H\u003csub\u003eA\u003c/sub\u003e and H\u003csub\u003eB\u003c/sub\u003e counterparts of QacA and Tet(L) are known to be important for their functions \u003csup\u003e\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e,\u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e33\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eStructural analysis of EmrB revealed a central aqueous cavity, CavEB\u003csub\u003eout\u003c/sub\u003e, located in the core of the membrane that opens toward the periplasmic space, indicating that EmrB adopts an outward-open state (Extended Data Fig.\u0026nbsp;10a). Entry into the cytoplasm from this cavity is restricted by interactions of side chains between helices H4 and H5 in the EBN domain and helices H10 and H11 in the EBC domain (Extended Data Fig.\u0026nbsp;10a). Within the central cavity, we identified a proton-titratable residue D29, located in motif D, which we hypothesized would be critical in the antiporter function of EmrB \u003csup\u003e\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e\u003c/sup\u003e. The corresponding residues of D34 in the TM1 of QacA and MfdA have previously been shown to be crucial for substrate recognition \u003csup\u003e\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e,\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e\u003c/sup\u003e. In our structure, D29 forms a salt bridge with R109 and hydrogen bonds with Q112 and N174, respectively (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003ed). All three of these residues are highly conserved among EmrB homologues (Extended Data Fig.\u0026nbsp;9b). The EmrB\u003csub\u003eD29N\u003c/sub\u003e or EmrB\u003csub\u003eR109A\u003c/sub\u003e mutation was found to abrogate bacterial resistance to nalidixic acid, nitroxoline and CCCP, supporting the hypothesis that these residues are involved in proton translocation during transport (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003ee). Interestingly, EmrA or EmrB alone is inactive, suggesting that the EmrAB complex is crucial for its function (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003ee).\u003c/p\u003e \u003cp\u003eParts of the 13th and 14th TM helices (H11 and H12) of EmrB protrude into the periplasm and form a structural unit named assembly subdomain B (AssB) (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003ea-c). The AssB is composed of an N-terminal helix X (H\u003csub\u003eX\u003c/sub\u003e) following H11, a C-terminal helix Z (H\u003csub\u003eZ\u003c/sub\u003e) preceding H12, and a horizontal helix Y (H\u003csub\u003eY\u003c/sub\u003e) parallel to the membrane, along with the loops connecting these helices. The most pronounced difference between EmrB and other MFS transporters is the unique AssB subdomain structure, which is key for pump assembly (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eb).\u003c/p\u003e\n\u003ch3\u003eInterfacial contacts between EmrA and EmrB\u003c/h3\u003e\n\u003cp\u003eInteractions of EmrB with other components of the tripartite assembly appear to be required for efflux activity, suggesting that the interactions allosterically modulate the activity of the MFS protein. The β-barrel domain of EmrA mediates interaction with EmrB, with the AssA subdomain of EmrA playing a key role through interactions with the periplasmic AssB subdomain of EmrB (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003ea). H\u003csub\u003eY\u003c/sub\u003e in the AssB subdomain penetrates the hydrophobic ring-1 in the AssA subdomain. Since H\u003csub\u003eY\u003c/sub\u003e is larger than the narrow entrance formed by hydrophobic ring-1, it becomes trapped above the hydrophobic ring-1, effectively blocking the narrow entrance (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eb, c; Extended Data Fig.\u0026nbsp;8b, c). The EmrA\u003csub\u003eL302Q/L303Q\u003c/sub\u003e mutation abrogated pump-mediated resistance to nalidixic acid, nitroxoline, and CCCP, indicating that the hydrophobic ring-1 plays an essential role in pump assembly (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003ed, e; Extended Data Fig.\u0026nbsp;8c). Below the hydrophobic ring-1, helices H\u003csub\u003eX\u003c/sub\u003e and H\u003csub\u003eZ\u003c/sub\u003e in the AssB subdomain tilt towards one side of CavEA, making contacts only with β-CL1 to β-CL3 (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eb).\u003c/p\u003e \u003cp\u003eIn addition, the cytoplasmic surface of the EmrA β-barrel domains directly engages with the periplasmic region of EmrB, and the N-terminal TM helices of EmrA insert into the membrane, stabilizing the interaction interface. The protruding loops in EmrB, positioned between helices H1\u0026minus;H2, H7\u0026minus;H8 and H9\u0026minus;H10, fit snugly into the concave pockets in the AssA subdomain. Helices H\u003csub\u003eX\u003c/sub\u003e and H\u003csub\u003eZ\u003c/sub\u003e of the AssB subdomain, along with other periplasmic regions between the transmembrane helices of EmrB, position along the side of the CavEA. The internal chamber formed is referred to as CavEA\u003csub\u003eout\u003c/sub\u003e, and its size is sufficient to accommodate all known EmrB substrates. The cavity CavEB\u003csub\u003eout\u003c/sub\u003e in EmrB opens to CavEA\u003csub\u003eout\u003c/sub\u003e, allowing the substrate in CavEB\u003csub\u003eout\u003c/sub\u003e to release into CavEA\u003csub\u003eout\u003c/sub\u003e (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003ec). Given the blockage of the narrow entrance by H\u003csub\u003eY\u003c/sub\u003e, the route of β-CL1PH in β-CL1 serves as the sole link between CavEA\u003csub\u003eout\u003c/sub\u003e and CavLP.\u003c/p\u003e \u003cp\u003eThe hydrophobic ring-2 in AssA assists in the formation of a sealed continuous channel between CavEB\u003csub\u003eout\u003c/sub\u003e and CavEA\u003csub\u003eout\u003c/sub\u003e by resting on the surface of the cellular membrane (Extended Data Fig.\u0026nbsp;8b, c). The EmrA\u003csub\u003eI313Q/V315T/V316T\u003c/sub\u003e mutation in hydrophobic ring-2 abrogated resistance to nalidixic acid, nitroxoline and CCCP, indicating that the hydrophobic ring-2 is critical for the pump assembly (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003ed, e; Extended Data Fig.\u0026nbsp;8c).\u003c/p\u003e \u003cdiv id=\"Sec8\" class=\"Section2\"\u003e \u003ch2\u003ePrediction of EmrB in its inward-open conformation\u003c/h2\u003e \u003cp\u003eEmrB has been suggested to adopt an inward-open state for substrate recognition and binding during transport, thereby forming a periplasmic gate \u003csup\u003e\u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e35\u003c/span\u003e\u003c/sup\u003e. However, in both our cryo-EM structure and all 5 models predicted with high-confidence scores by AlphaFold3, EmrB is consistently displayed in an outward-open state (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003ea; Extended Data Fig.\u0026nbsp;11a, b) \u003csup\u003e\u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e\u003c/sup\u003e. In an attempt to capture EmrB in its inward-open state, we utilised the multimer feature of AlphaFold3 to generate a prediction of EmrB in complex with six molecules of EmrA. In all 5 of the models generated by AlphaFold, we were able to observe EmrB in a distinct conformation (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eb; Extended Data Fig.\u0026nbsp;11c, d). Based on the disappearance of CavEB\u003csub\u003eout\u003c/sub\u003e and the presence of a new and enlarged cytosolic cavity (named CavEB\u003csub\u003ein\u003c/sub\u003e), we hypothesized that AlphaFold had been able to capture the inward-open state of EmrB previously unobservable in the experimental data set (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003ec) \u003csup\u003e\u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e35\u003c/span\u003e\u003c/sup\u003e. Using a homology model of EmrB in the inward-open state in conjunction with these predictions, we identified several residues which we believed would be important in the process of translocation. Residues N62, V147, I148 and I288 are all situated along the wall of the cytosolic-facing cavity (Extended Data Fig.\u0026nbsp;10b). The corresponding residues in MdfA are involved in drug recognition (Extended Data Fig.\u0026nbsp;12) \u003csup\u003e\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e\u003c/sup\u003e. In support of AlphaFold, mutations of N62, V147, and I288 indeed reduced pump-mediated resistance to nitroxoline and nalidixic acid. I148 was crucial for resistance to CCCP, whereas it appeared dispensable for nitroxoline and nalidixic acid (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003ef, g). Nalidixic acid/H⁺ antiport activity was detected in everted membrane vesicles expressing wild-type EmrB and its mutants. Wild-type EmrB exhibited significant H⁺ movement, whereas the EmrB\u003csub\u003eD29N\u003c/sub\u003e and EmrB\u003csub\u003eV147A\u003c/sub\u003e mutants showed reduced H⁺ movement (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eh). Altogether, these results support the predicted model for the inward-open state and demonstrating the importance of specific residues within EmrB in exhibiting poly-specific drug recognition during the transport process.\u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003eProposed structural transitions of EmrB\u003c/h3\u003e\n\u003cp\u003eUpon close inspection of the structures of EmrB in different states, we detected significant changes in the chemical environments of residues already revealed to be important in our mutagenesis studies. For example, the proposed proton-carrier D29 was shown to establish 3 hydrogen bonds in the outward-open state in comparison to just 2 hydrogen bonds in the inward-open state (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003ed). To provide further evidence that AlphaFold3 had indeed captured the inward-open state and to try gain insight into conformational changes within EmrB, we employed a modified version of AlphaFold2, known as AF-Cluster, to obtain predictions of the various states which EmrB could adopt \u003csup\u003e\u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e37\u003c/span\u003e\u003c/sup\u003e. By grouping the multiple-sequence alignments based on sequence similarity within the MFS family and subsequently performing AlphaFold2 predictions, we captured a plethora of structures representative of the transition between the outward-open to inward-open state of EmrB (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003ee). Interestingly, 75% of these high-confidence predictions could be classified into structures representative of the inward-open states, providing further evidence of this conformation being adopted by EmrB. Furthermore, these predictions also provided insight into the conformational dynamics that EmrB is likely to undergo in shuttling substrates, analogous to the transitions proposed for other members of the MFS family \u003csup\u003e\u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e35\u003c/span\u003e\u003c/sup\u003e. The helices within the EBC domain appear to remain static, whereas helices H1 \u0026ndash; H6 in the EBN domain are suspected to be highly dynamic. The helices appear to pivot at a central point, in support of a \u0026lsquo;rocker-switch\u0026rsquo; mechanism of transport for EmrB, common to other MFS antiporters \u003csup\u003e\u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e38\u003c/span\u003e\u003c/sup\u003e. Altogether, these results provide evidence of an inward-open conformation of EmrB, as well as insights into the structural transitions of EmrB that are important in the energy-dependent flux of antibiotics by EmrAB-TolC. The interactions of EmrA-EmrB change in the structural transitions of EmrB, and may provide the path of allosteric communication of these components.\u003c/p\u003e\n\u003ch3\u003eInteractions between EmrA and TolC\u003c/h3\u003e\n\u003cp\u003eA short helix-turn-helix motif located within the α-helical hairpin domain of EmrA mediates tip-to‐tip interactions with the periplasmic end of TolC, similar to the interactions observed between the homologous MacA and TolC (Extended Data Fig.\u0026nbsp;13) \u003csup\u003e\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e\u003c/sup\u003e. Engagement between TolC and EmrA directly opens the periplasmic aperture of TolC, creating a continuous nanochannel with a width of 25\u0026minus;30 \u0026Aring;, termed NanoTC. TolC is predicted to remain open during the pumping process, as observed previously for RND-type AcrAB-TolC and MexAB-OprM pumps \u003csup\u003e\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e,\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e,\u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e39\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e \u003cdiv id=\"Sec11\" class=\"Section2\"\u003e \u003ch2\u003eOne-step drug transport mechanism of the EmrAB-TolC pump\u003c/h2\u003e \u003cp\u003eThe structure of the EmrAB-TolC tripartite pump represents the resting apo state structure. The combined data with the Alphafold3 prediction structures suggest that it employs a one-step drug transport mechanism across the entire cell envelope (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003e). EmrA and EmrB possess unique AssA and AssB subdomains, respectively (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eb; Extended Data Fig.\u0026nbsp;8b, c). The six β-CLs in AssA vary in conformation, one of which develops a path (β-CL1PH) linking the chambers CavLP and CavEA\u003csub\u003eout\u003c/sub\u003e in the lipoyl and β-barrel domains (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eb, c; Extended Data Fig.\u0026nbsp;8a-c). Residues L302 and L303 in the β-CL hexamer form a hydrophobic ring-1, onto which the AssB anchors (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e; Extended Data Fig.\u0026nbsp;8b, c). EmrA lacks the MP domain, which enables EmrA and EmrB to form a secure conduit and its hydrophobic ring-2 strengthens this conduit (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003ed; Extended Data Fig.\u0026nbsp;8b, c). The structure of the EmrAB-TolC pump delineates the complete transport pathway from the central cavity CavEB\u003csub\u003eout\u003c/sub\u003e of EmrB to the cell exterior (Extended Data Fig.\u0026nbsp;14). As a homologue of QacA and MdfA \u003csup\u003e\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e,\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e,\u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e40\u003c/span\u003e\u003c/sup\u003e, EmrB in the IM could adopt an inward-open state to facilitate substrate recognition and binding from the inner leaflet of the IM and cytoplasm, as supported by AlphaFold3 predictions, homology modelling and mutagenesis studies (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003ef; Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003e; Extended Data Fig.\u0026nbsp;10b). Subsequently, the substrate is transported to CavEA\u003csub\u003eout\u003c/sub\u003e via an alternating access mechanism, characteristic of MFS transporters (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003e) \u003csup\u003e\u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e35\u003c/span\u003e\u003c/sup\u003e. The protonation of D29 transforms EmrB to an inward-open state again. The periplasmic regions between the transmembrane helices of EmrB shift toward the central axis, which partially occupy the space in CavEA and cause CavEA\u003csub\u003eout\u003c/sub\u003e to shrink. The compaction of CavEA\u003csub\u003eout\u003c/sub\u003e drives substrates toward the β-CL1PH-CavLP pathway to the nanochannel formed by EmrA and TolC. This contraction of CavEA has been proposed in the AlphaFold3-predicted EmrAB structure (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003ec). The transport pathway involves several essential elements, namely, CavEB\u003csub\u003ein\u003c/sub\u003e, CavEB\u003csub\u003eout\u003c/sub\u003e, CavEA\u003csub\u003eout\u003c/sub\u003e, β-CL1PH, CavLP, NanoHP, and NanoTC (Extended Data Fig.\u0026nbsp;10a, b; Extended Data Fig.\u0026nbsp;14). These components form a secure conduit for substrate release to the cell exterior (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003e). The gating helices in EmrB allow substrates to flow against concentration gradients and prevent backflow during operation (Extended Data Fig.\u0026nbsp;10a, b), resulting in the export of drugs across the IM and OM to the extracellular milieu.\u003c/p\u003e \u003c/div\u003e"},{"header":"Discussion","content":"\u003cp\u003eAntibiotic resistance of pathogenic bacteria is a growing clinical problem. Drug efflux pumps in the bacterial cell envelope play important roles in multidrug resistance. In the current study, the \u003cem\u003eE. coli\u003c/em\u003e EmrAB-TolC pump was shown to confer resistance to several different antimicrobial drugs, including nalidixic acid, nitroxoline and CCCP (Extended Data Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003ea). The EmrAB-TolC structure identifies key residues that contribute to such multi-drug resistance (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003ed; Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003ef). The conserved hydrophobic ring-1 of EmrA is crucial for the assembly and functioning of the pump and is accessible to the external environment via the open nanochannel in the full pump assembly (Extended Data Fig.\u0026nbsp;9a; Extended Data Fig.\u0026nbsp;14); thus, it represents a promising target for the development of drugs that inhibit this class of pump.\u003c/p\u003e \u003cp\u003eAt present, six families of efflux pumps are known to be involved in drug transport: primary transporters of the ATP-binding cassette (ABC) family, and secondary transporters including the major facilitator (MFS), resistance/nodulation/cell division (RND), small multidrug resistance (SMR), multidrug and toxic compounds extrusion (MATE), and proteobacterial antimicrobial compound efflux (PACE) families. In Gram-negative bacteria, all these drug transporters are located in the IM. The members of five families of drug transporters (MFS, SMR, MATE, PACE and ABC) usually function as independent units to translocate substrates across the IM bilayer \u003csup\u003e\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e\u003c/sup\u003e. Drugs on the periplasmic side must be further translocated across the OM. Structures of RND-type AcrAB-TolC and ABC-type MacAB-TolC reveal that the IM components AcrB and MacB collect substrates from the outer leaflet of the IM and periplasm, and efflux them to the extracellular milieu through the tripartite assembly (Extended Data Fig.\u0026nbsp;15a, b) \u003csup\u003e\u003cspan additionalcitationids=\"CR6\" citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e\u003c/sup\u003e. It is likely that the five transporter systems cooperate with the double membrane-spanning RND-type or ABC-type tripartite pumps to deliver substrates across the OM to the cell exterior. Thus, these tripartite efflux pumps employ a two-step mechanism of periplasmic substrate capture and transport (Extended Data Fig.\u0026nbsp;15a, b) \u003csup\u003e\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e,\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e\u003c/sup\u003e. On the other hand, MFS-type tripartite pumps appear to utilize a one-step mechanism to transport drugs across the entire cell envelope, bypassing the periplasm (Extended Data Fig.\u0026nbsp;15c). The closely related T1SS assemblies, such as HlyDB-TolC, are believed to employ similar processes for the secretion of protein substrates\u003csup\u003e\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e\u003c/sup\u003e. The HlyDB-TolC pump could form a tripartite assembly. Similar to EmrA, the adapter protein HlyD lacks a MP domain (Extended Data Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003e)\u003csup\u003e\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e,\u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e41\u003c/span\u003e\u003c/sup\u003e, which may enable HlyD and HlyB to form a secure conduit. The ABC family transporter HlyB could drive the secretion substrate HlyA across double membranes of the cells via an alternating access mechanism. We note that RND-type tripartite pumps can employ a one-step transport mechanism across the outer membrane for certain substrates, such as beta-lactam antibiotics \u003csup\u003e\u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e42\u003c/span\u003e\u003c/sup\u003e, but it is rather different mechanistically from the EmrAB-TolC pump.\u003c/p\u003e \u003cp\u003eThe EmrAB-TolC, AcrAB-TolC and MacAB-TolC pumps all utilize the TolC OM channel, and have structurally similar periplasmic adaptor proteins that form hexametric channels, but their IM components are derived from distinct transporter families. The periplasmic adaptor proteins have evolved uniquely to complement the specific IM transporters. As a result, each type of tripartite pump utilizes a unique pump assembly and gating mechanism that contributes to its function. The ability of these proteins to recognize and transport various substrates largely depends on the specific domains within their IM transporter components, and interactions of the components can allosterically modulate the activity of the inner membrane component. Here, we observe extensive interactions between EmrA and EmrB that change with conformational switch. These interactions are proposed to modulate the activity of EmrB so that it does not function as a stand-alone transporter, as occurs for example in the well-studied MFS proteins QacA and MfdA. Understanding the similarities and differences between tripartite multidrug efflux pumps and the basis for their allostery is critical for developing strategies to combat multidrug resistance, a global health challenge.\u003c/p\u003e "},{"header":"Materials And Methods","content":"\u003cp\u003e\u003cstrong\u003eConstruction of EmrAB-TolC complex overexpression vectors\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe \u003cem\u003eBRIL\u003c/em\u003e gene was synthesized by AZENTA (Suzhou, China) and amplified using the brilpet-F and bril-link-R primers. The \u003cem\u003eemrA\u003c/em\u003e,\u003cem\u003e\u0026nbsp;emrB\u003c/em\u003e,\u003cem\u003e\u0026nbsp;tolC\u003c/em\u003e, and \u003cem\u003eacrA\u003c/em\u003e genes were amplified from genomic DNA of \u003cem\u003eE. coli\u003c/em\u003e K-12 strain MG1655. The \u003cem\u003eemrB\u003c/em\u003e (37-1494 bp) genefragmentwas amplified using primers emrB-link-F and emrBGS-R. The full-length \u003cem\u003eemrA\u003c/em\u003e gene was first amplified using primers emrAGS-F and emrApet-R; the pETDuet-1 plasmid was linearized using primer pair Petduet_F and Petduet_R; and the \u003cem\u003eBRIL\u003c/em\u003e, \u003cem\u003eemrB\u003c/em\u003e, and\u003cem\u003e\u0026nbsp;emrA\u003c/em\u003e DNA fragments were inserted into the pETDuet-1 vector using NovoRec plus One step PCR Cloning Kit (Novorprotein, Shanghai, China), resulting in pETDuet-\u003cem\u003eBRIL-emrB-polyGlySer-emrA-6His\u003c/em\u003e. The \u003cem\u003eALFA-tag\u003c/em\u003e sequence was subsequently inserted into the 3\u0026prime; end of \u003cem\u003eemrB\u003c/em\u003e by site-directed mutagenesis, generating pETDuet-\u003cem\u003eBRIL-emrB-ALFA-polyGlySer-emrA-6His\u003c/em\u003e (abbreviated as pET_BA).\u003c/p\u003e\n\u003cp\u003eUsing primers TolCinf_F and TolCFLAGXhoI_R, the \u003cem\u003etolC\u003c/em\u003e gene was amplified, and the product served as a template for a second amplification using primers TolCinf_F and TolCFLAG_inf_R. The resulting \u003cem\u003etolC-FLAG\u003c/em\u003e DNA fragment was inserted into pRSFDuet-1 (digested with \u003cem\u003eNde\u003c/em\u003eI) using the\u0026nbsp;in-fusion ligation method of the NovoRec plus One step PCR Cloning Kit (Novorprotein, Shanghai, China), thereby generating pRSFDuet-\u003cem\u003etolC-FLAG\u003c/em\u003e. The \u003cem\u003eemrA\u003c/em\u003e DNA fragment was amplified using primers emrAinsert_F and emrAinsert_R and inserted into the pRSFDuet-\u003cem\u003etolC-FLAG\u003c/em\u003e vector using the in-fusion cloning method (Novorprotein, Shanghai, China), resulting in pRSFDuet-\u003cem\u003eemrA-tolC-FLAG\u003c/em\u003e (abbreviated as pRSF_AC).\u003c/p\u003e\n\u003cp\u003ePrimer pair \u003cem\u003e\u0026Delta;\u003c/em\u003eemrAinsert_F and emrAinsert_R was used to amplify the truncated \u003cem\u003e\u0026Delta;emrA\u003c/em\u003e gene (142-1173 bp), and the \u003cem\u003e\u0026Delta;emrA\u003c/em\u003e DNA fragment was inserted into the pRSFDuet-\u003cem\u003etolC-FLAG\u003c/em\u003e vector using the In-Fusion cloning method (Novorprotien, Shanghai, China), generating pRSFDuet-\u003cem\u003e\u0026Delta;emrA-tolC-FLAG\u003c/em\u003e. The N-terminal 81 bp region of \u003cem\u003eacrA\u003c/em\u003e (\u003cem\u003eacrAsignal\u003c/em\u003e) was amplified using the acrAs_F and acrAs_R primer pair, and the resulting \u003cem\u003eacrAsignal\u003c/em\u003e DNA fragment was inserted into the 5\u0026apos; end of \u003cem\u003e\u0026Delta;emrA\u003c/em\u003e in pRSFDuet-\u003cem\u003e\u0026Delta;emrA-tolC-FLAG\u003c/em\u003e using the In-fusion cloning method (Novorprotien, Shanghai, China), generating pRSFDuet-\u003cem\u003eacrAsignal-\u0026Delta;emrA-tolC-FLAG\u003c/em\u003e (abbreviated as pRSF_\u003cem\u003e\u0026Delta;\u003c/em\u003eAC).\u003c/p\u003e\n\u003cp\u003eThe \u003cem\u003eemrAB\u003c/em\u003e genes were amplified from genomic DNA of \u003cem\u003eE. coli\u003c/em\u003e K12 strain MG1655 using primers EmrAB_F and EmrAB_R, the pETDuet-1 plasmid was amplified using primers Petduet_V_F and Petduet_V_R, and the \u003cem\u003eemrAB\u003c/em\u003e fragment was inserted into pETDuet-1 using an In-Fusion Ligation Kit, resulting in pETDuet-\u003cem\u003eemrAB\u003c/em\u003e.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eThe site-directed mutants EmrAB\u003csub\u003eD29N\u003c/sub\u003e and EmrAB\u003csub\u003eR109A\u003c/sub\u003e were produced using pETDuet-\u003cem\u003eemrAB\u003c/em\u003e as a template with the primer pairs emrB\u003csub\u003eD29N\u003c/sub\u003e_F/emrB\u003csub\u003eD29N\u003c/sub\u003e_R and emrB\u003csub\u003eR109A\u003c/sub\u003e_F/emrB\u003csub\u003eR109A\u003c/sub\u003e_R, resulting in pETDuet-\u003cem\u003eemrAB\u003csub\u003eD29N\u0026nbsp;\u003c/sub\u003e\u003c/em\u003eand pETDuet-\u003cem\u003eemrAB\u003csub\u003eR109A\u003c/sub\u003e\u003c/em\u003e.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eThe site-directed mutants EmrAB\u003csub\u003eN62A\u003c/sub\u003e, EmrAB\u003csub\u003eV147A\u003c/sub\u003e, EmrAB\u003csub\u003eV148A\u003c/sub\u003e, and EmrAB\u003csub\u003eI288A\u003c/sub\u003e were created by using pETDuet-\u003cem\u003eemrAB\u0026nbsp;\u003c/em\u003eas a template with the primer pairs emrB\u003csub\u003eN62A\u003c/sub\u003e_F/emrB\u003csub\u003eN62A\u003c/sub\u003e_R, emrB\u003csub\u003eV147A\u003c/sub\u003e_F/emrB\u003csub\u003eV147A\u003c/sub\u003e_R, emrB\u003csub\u003eV148A\u003c/sub\u003e_F/emrB\u003csub\u003eV148A\u003c/sub\u003e_R, and emrB\u003csub\u003eI288A\u003c/sub\u003e_F/emrB\u003csub\u003eI288A\u003c/sub\u003e_R, resulting in pETDuet-\u003cem\u003eemrAB\u003csub\u003eN62A\u003c/sub\u003e\u003c/em\u003e, pETDuet-\u003cem\u003eemrAB\u003csub\u003eV147A\u003c/sub\u003e\u003c/em\u003e, pETDuet-\u003cem\u003eemrAB\u003csub\u003eV148A\u003c/sub\u003e\u003c/em\u003e, and pETDuet-\u003cem\u003eemrAB\u003csub\u003eI288A\u003c/sub\u003e\u003c/em\u003e. \u0026nbsp;\u003c/p\u003e\n\u003cp\u003eThe site-directed mutants EmrA\u003csub\u003eL302Q/L303Q\u003c/sub\u003eB and EmrA\u003csub\u003eI313Q/V315T/V316T\u003c/sub\u003eB were produced using pETDuet-\u003cem\u003eemrAB\u0026nbsp;\u003c/em\u003eas a template with the primer pairs emrA\u003csub\u003eL302Q/L303Q\u003c/sub\u003e_F/emrA\u003csub\u003eL302Q/L303Q\u003c/sub\u003e_R and emrA\u003csub\u003eI313Q/V315T/V316T\u003c/sub\u003e_F/ emrA\u003csub\u003eI313Q/V315T/V316T\u003c/sub\u003e_R, resulting in plasmids pETDuet-\u003cem\u003eemrA\u003csub\u003eL302Q/L303Q\u003c/sub\u003eB\u003c/em\u003e and pETDuet-\u003cem\u003eemrA\u003csub\u003eI313Q/V315T/V316T\u003c/sub\u003eB\u003c/em\u003e.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eThe enzyme used to introduce the mutations above-mentioned was purchased from CloneAmp HiFi PCR Premix (Clontech, Germany). Then the modified plasmids above-mentioned were amplified and extracted using SPARKeasy Mini Plasmid Ultra-Fast Kit (Shandong Sparkjade Biotechnology Co.,Ltd.).\u003c/p\u003e\n\u003cp\u003ePrimers used to generate all constructs are listed in Extended Data Table 1.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eOverexpression and purification of \u003cem\u003eE. coli\u003c/em\u003e EmrAB-TolC\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe plasmids pET_BA and pRSF_\u003cem\u003e\u0026Delta;\u003c/em\u003eAC were used to co-express and purify EmrAB-TolC pump-EA, which contains the EmrB-EmrA fusion protein, TolC and a modified free EmrA with the N-terminal residues 1-47 replaced by residues 1-27 of AcrA. The 6xHis-Tag at the C-terminus of the EmrB-EmrA fusion protein was initially used to pull down the co-expressed pump components. Next, the FLAG-Tag at the C-terminus of TolC was used to isolate the fully assembled EmrAB-TolC pump. BRIL and ALFA-Tag\u0026nbsp;was used to\u0026nbsp;stabilize EmrB in the complex.\u0026nbsp;The \u003cem\u003eE. coli\u003c/em\u003e C43 (DE3) \u003cem\u003e\u0026Delta;acrAB\u003c/em\u003e strain was transformed with plasmids pET_BA and pRSF_\u003cem\u003e\u0026Delta;\u003c/em\u003eAC. From an agar plate containing appropriate antibiotics, a single colony was picked and inoculated into 20 mL of LB medium supplemented with 50 \u0026mu;g mL\u003csup\u003e-1\u003c/sup\u003e kanamycin and 100 \u0026mu;g mL\u003csup\u003e-1\u003c/sup\u003e carbenicillin in a 50 mL centrifuge tube. The culture was incubated at 37\u0026deg;C with shaking at 220 rpm for 4-5 h.\u003c/p\u003e\n\u003cp\u003eA 10 mL culture aliquot was then transferred to 1 L of 2\u0026acute;YT medium containing antibiotics in a 2 L baffled flask. The same incubation conditions were applied. Cells were cultured until a density (absorbance at 600 nm, A\u003csub\u003e600\u003c/sub\u003e) of 0.5-0.6, induced with 0.1 mM isopropyl-\u0026beta;-D-thiogalactoside (IPTG), the temperature was reduced to 20\u0026deg;C, and culturing was continued overnight.\u003c/p\u003e\n\u003cp\u003eThe cells were harvested by centrifugation, and the cell pellet from 6 L of culture was resuspended in 200 mL of lysis buffer containing 20 mM Tris (pH 8.0) and 300 mM NaCl. Afterwards, 2 mL of 100\u0026acute;EDTA-Free Protease Inhibitor Cocktail (APExBIO, USA) was added to a final concentration of 1\u0026acute;. Lysozyme and DNase I were added to final concentrations of 5 mg/mL and 5 U/mL, respectively. The cell mixture was incubated at 4\u0026deg;C for 1 h before high-pressure homogenization via three passages at 15,000 psi at 4\u0026deg;C. The lysate was centrifuged to remove cell debris, and the supernatant was ultracentrifuged to pellet the cellular membranes.\u003c/p\u003e\n\u003cp\u003eLysis buffer was used to resuspend the membrane pellet. EDTA-free protease inhibitor cocktail (100\u0026acute;) was added to a final concentration of 1\u0026acute;, \u003cem\u003en\u003c/em\u003e-dodecyl-\u003cem\u003e\u0026beta;\u003c/em\u003e-D-maltoside (DDM) was added to a final concentration of 1.5%, and the mixture was gently stirred at 4\u0026deg;C for 3 h. The mixture was ultracentrifuged. Imidazole was added to the supernatant to a final concentration of 20 mM, and the mixture was applied to a 1 mL HiTrap Chelating column (Cytiva, USA) charged with Ni\u003csup\u003e2+.\u003c/sup\u003e\u003c/p\u003e\n\u003cp\u003eThe column was washed with lysis buffer containing 0.05% DDM and 50 mM imidazole, followed by lysis buffer supplemented with 0.01% DDM. Peptidisc \u003csup\u003e43,44\u003c/sup\u003e (1 mL) at a concentration of 5 mg/mL in lysis buffer was injected onto the column and incubated at 4\u0026deg;C for 1 h. The column was washed with 10 mL of lysis buffer containing 1 mg/mL peptide disc. Elution of the 6\u0026acute;His-tagged EmrAB-TolC protein complex was achieved with lysis buffer containing 300 mM imidazole. Buffer exchange of the eluate into Buffer-I, which consisted of 20 mM Tris (pH 7.5) and 150 mM NaCl, was conducted using a HiTrap Desalting column (Cytiva, USA). The 6\u0026acute;His-tagged EmrAB-TolC complex was further purified using ANTI-FLAG M2 affinity resin (GenScript, Nanjing, China). The resin was prepared by sequential washing with glycine HCl (pH 3.5) and Buffer-I. The protein solution and resin were mixed, gently rotated for 1 h at 4\u0026deg;C, loaded into a chromatography column, and washed with Buffer-I. The mixture was resuspended in 1 mg/mL FLAG-peptide in Buffer-I, incubated for 30 min, centrifuged, and the supernatant was loaded onto a mini chromatography column to remove residual resin. This step was repeated three times. Fractions containing EmrAB-TolC were pooled, concentrated using a centrifugal filter unit (Merck, Germany) with a molecular weight cut-off of 100 kDa, flash-frozen in liquid nitrogen and stored at -80\u0026deg;C.\u003c/p\u003e\n\u003cp\u003eThe plasmids pET_BA and pRSF_AC were used to co-express and purify the EmrAB-TolC pump-FA containing the EmrB-EmrA fusion protein, TolC, and wild-type free EmrA using the same procedure employed for the pump-EA.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCell growth for the antibiotic sensitivity assay\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eSingle colonies of \u003cem\u003eE. coli\u0026nbsp;\u003c/em\u003eC43(DE3)\u003cem\u003e\u0026nbsp;\u0026Delta;acrAB\u003c/em\u003e harboring the plasmids pETDuet/pRSFDuet, pET_BA/pRSF_AC, pET_BA/pRSF_\u003cem\u003e\u0026Delta;\u003c/em\u003eAC, pETDuet-\u003cem\u003eemrAB\u003c/em\u003e/pRSFDuet-\u003cem\u003etolC-FLAG\u003c/em\u003e, pETDuet-\u003cem\u003eemrAB\u003csub\u003eD29N\u003c/sub\u003e\u003c/em\u003e/pRSFDuet-\u003cem\u003etolC-FLAG\u003c/em\u003e, pETDuet-\u003cem\u003eemrAB\u003csub\u003eR109A\u003c/sub\u003e\u003c/em\u003e/pRSFDuet-\u003cem\u003etolC-FLAG\u003c/em\u003e, pETDuet-\u003cem\u003eemrA\u003csub\u003eL30Q/L303Q\u003c/sub\u003eB\u003c/em\u003e/pRSFDuet-\u003cem\u003etolC-FLAG\u003c/em\u003e, pETDuet-\u003cem\u003eemrA\u003csub\u003eI313A/V315A/V316A\u003c/sub\u003eB\u003c/em\u003e/pRSFDuet-\u003cem\u003etolC-FLAG\u003c/em\u003e, pETDuet-\u003cem\u003eemrA\u003c/em\u003e\u003cem\u003eB\u003csub\u003eN62A\u003c/sub\u003e\u003c/em\u003e/pRSFDuet-\u003cem\u003etolC-FLAG\u003c/em\u003e, pETDuet-\u003cem\u003eemrA\u003c/em\u003e\u003cem\u003eB\u003csub\u003eV147A\u003c/sub\u003e\u003c/em\u003e/pRSFDuet-\u003cem\u003etolC-FLAG\u003c/em\u003e, pETDuet-\u003cem\u003eemrA\u003c/em\u003e\u003cem\u003eB\u003csub\u003eV148A\u003c/sub\u003e\u003c/em\u003e/pRSFDuet-\u003cem\u003etolC-FLAG\u003c/em\u003e, or pETDuet-\u003cem\u003eemrA\u003c/em\u003e\u003cem\u003eB\u003csub\u003eI288A\u003c/sub\u003e\u003c/em\u003e/pRSFDuet-\u003cem\u003etolC-FLAG\u003c/em\u003e were grown in fresh LB supplemented with 50 \u0026micro;g/mL kanamycin and 100 \u0026micro;g/mL ampicillin at 37℃\u0026nbsp;with shaking at 220 rpm. When the optical density at 600 nm (OD\u003csub\u003e600\u003c/sub\u003e) reached 0.5\u0026ndash;0.6, expression was induced with 100 \u0026micro;M IPTG, the temperature was reduced to 22\u0026deg;C, and the incubation was continued for 1 h.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eThe cells harboring pET_BA/pRSF_AC or pET_BA/pRSF_\u003cem\u003e\u0026Delta;\u003c/em\u003eAC were harvested by centrifugation, after which the cell pellets were resuspended in fresh medium to an OD\u003csub\u003e600\u003c/sub\u003e of 0.1 and treated with 0.8\u0026nbsp;mg/mL nitroxoline and 30 \u0026micro;M IPTG. Cell growth was monitored over time by measuring the OD\u003csub\u003e600\u003c/sub\u003e at 37\u0026deg;C in a Spark microplate reader (Tecan, Austria).\u003c/p\u003e\n\u003cp\u003eThe cells harboring the remaining constructs were harvested by centrifugation, resuspended in fresh medium to an OD\u003csub\u003e600\u003c/sub\u003e of 0.5, and 10-fold serially diluted. Then 5 \u0026micro;L of each culture was plated on LB agar containing 60 \u0026micro;M IPTG in the presence of 1 \u0026micro;g/mL nalidixic acid, 5 \u0026micro;g/mL nitroxoline, or 1 \u0026micro;g/mL CCCP. \u0026nbsp;As a control, 5 \u0026mu;L of all suspensions were plated on LB agar without antibiotics. Cell growth was monitored after culture at 37\u0026deg;C overnight. The assay was performed three times independently.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eDrug-proton antiport assay\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe assay was conducted as described previously \u003csup\u003e40,45-47\u003c/sup\u003e. Inside-out (everted) membrane vesicles were prepared from C43 (DE3) \u0026Delta;acrAB cells expressing the EmrB variants. Briefly, bacterial cells were cultured in 250 ml LB medium at 37\u0026deg;C to an OD600 of 0.5, then induced for protein expression by adding 0.2 mM IPTG at 37\u0026deg;C for 2 hours to an OD600 of 1.5. The cells were harvested by centrifugation at 7,500 x g for 10 minutes, resuspended in 20 ml of membrane vesicle buffer A (200 mM Tris-HCl, pH 8.0, 2 mM EDTA, 30% sucrose) at room temperature, and incubated with 100 \u0026mu;g/ml lysozyme for 5-10 minutes. After adding 4 mM MgCl\u003csub\u003e2\u003c/sub\u003e and DNase\u0026nbsp;I, the cell spheroplasts were pelleted at 7,500 x g for 10 minutes, then resuspended in 25 ml of ice-cold membrane vesicle buffer B (100 mM HEPES-KOH, pH 7.5, 100 mM KCl, 10 mM MgCl\u003csub\u003e2\u003c/sub\u003e), supplemented with 2 mM DTT and 250 \u0026mu;L protease inhibitor\u0026nbsp;(K1007; APExBIO). The spheroplasts\u0026nbsp;were\u0026nbsp;lysed by five passages through a homogenizer (ATS, AH-NANO) at\u0026nbsp;400-500 bar. The unbroken cells and debris were pelleted at 7,500 x g for 10 minutes, and the membranes in the supernatant were pelleted at 150,000 x g for 40 minutes at 4\u0026deg;C in a Ti70 rotor (Beckman Coulter). The membrane pellet was resuspended in membrane vesicle buffer B by gently squirting the buffer toward the pellet with a 200 \u0026mu;L tip. The vesicles were then centrifuged at 3,500g at 4 \u0026deg;C for 30 s to remove the aggregates.\u003c/p\u003e\n\u003cp\u003eTo ensure equal quantities of everted membrane vesicles were used for each measurement, vesicles containing 200 \u0026mu;g total membrane proteins were added to 2 ml of pre-warmed (30\u0026deg;C) buffer\u0026nbsp;（20 mM Tris, 140 mM KCl, 10 mM MgCl\u003csub\u003e2\u003c/sub\u003e, pH 6.0）\u0026nbsp;and 1 \u0026mu;M 9-Amino-6-chloro-2-methoxyacridine (ACMA). Membranes were equilibrated for 4 minutes prior to fluorescence measurement with excitation at 409 nm and emission at 474 nm using an FluoroMax-4 spectrofluorometer (HORIBA Scientific). Before adding the substrates, 2 mM lactate was used and equilibrated for 3 minutes to energize the membrane, quenching the ACMA fluorescence. Fluorescence dequenching was observed upon adding 250 \u0026mu;M nalidixic acid, likely due to the extrusion of H\u003csup\u003e+\u003c/sup\u003e by antiporters translocating the drugs into the vesicles. To dissipate the transmembrane H\u003csup\u003e+\u003c/sup\u003e gradient, 5 mM NH\u003csub\u003e4\u003c/sub\u003eCl was added. The fluorescence changes induced by lactate and NH\u003csub\u003e4\u003c/sub\u003eCl further confirmed that similar amounts of everted vesicles of comparable quality were used in each measurement.\u003c/p\u003e\n\u003cp\u003e6xHis-tag was added to the N-terminus of EmrA or C-terminus of EmrB of EmrAB-TolC pump, respectively. The same procedure was used to prepare everted membrane vesicles for Western blot assays of EmrA/EmrB variants (Fig. 4e; Fig. 5g).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eElectron microscopy data collection\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eFor the EmrAB-TolC pump-EA cryo-EM assays, 3.5 \u0026mu;L aliquots of purified protein in peptidisc (protein concentration = 0.7 mg/mL) were added to glow-discharged holey carbon grids (Quantifoil Au R1.2/1.3, 300 mesh; Quantifoil Micro Tools GmbH). Blotting was performed with filter paper for 3.5 s to remove excess sample, and a Vitrobot Mark IV instrument (Thermo Fisher Scientific) was used for rapid freezing in a liquid ethane slush. A Titan Krios electron microscope (Thermo Fisher Scientific) operating at 300 kV coupled with a SerialEM and a Gatan K3-Summit detector (Gatan, Inc.) operating in super-resolution counting mode \u003csup\u003e48\u003c/sup\u003e were used to automatically collect zero-energy-loss images of frozen and hydrated grids. Using a slit width of 20 eV, a GIF-Quantum energy filter (Gatan) was applied to exclude inelastically scattered electrons. Using a dose rate of\u0026nbsp;\u0026sim;15.15 electrons \u0026Aring;\u003csup\u003e-2\u003c/sup\u003e s\u003csup\u003e-1\u003c/sup\u003e (~18 electrons pixel\u003csup\u003e-1\u003c/sup\u003e s\u003csup\u003e-1\u003c/sup\u003e) at an adjusted magnification of\u0026nbsp;\u0026acute;45,871.6 (yielding a pixel size of 1.09 \u0026Aring; at the sample level) and a total dose of\u0026nbsp;\u0026sim;60 electrons \u0026Aring;\u003csup\u003e\u0026minus;2\u003c/sup\u003e at the sample, 60 movie frames were recorded. The final dataset comprised 4334 movie stacks with defocus values between -1.0 and -2.5 \u0026mu;m.\u003c/p\u003e\n\u003cp\u003eFor the EmrAB-TolC pump-FA, a holey carbon grid (Quantifoil Au R1.2/1.3, 300 mesh) was overlayed with a homemade graphene monolayer and cleaned with UV/ozone at room temperature for 10 min using a Gatan SOLARUS (950) Plasma Cleaning System (Gatan, Inc.), ensuring hydrophilicity of the graphene grid. Next, 3.5 \u0026mu;L aliquots of the purified protein sample from the peptidisc (0.1 mg/mL concentration) were applied to the grid and incubated for 30 s. Excess sample was removed by blotting with filter paper for 3.5 s, followed by rapid freezing in liquid ethane slush using a Vitrobot Mark IV instrument. Frozen-hydrated EmrAB-TolC particles were subjected to automatic data collection using a Titan Krios electron microscope at 300 kV with SerialEM and a Gatan K3-Summit direct electron detector running in super-resolution counting mode at an adjusted magnification of 47,169.9, equating to a measured physical pixel size of 1.06 \u0026Aring; and a dose rate of\u0026nbsp;\u0026sim;16 electrons \u0026Aring;\u003csup\u003e-2\u003c/sup\u003e s\u003csup\u003e-1\u0026nbsp;\u003c/sup\u003e(\u0026sim;18 electrons pixel\u003csup\u003e-1\u003c/sup\u003e s\u003csup\u003e-1\u003c/sup\u003e). Exposures lasting 3.533 s were split into 60 movie frames, resulting in an accumulated dose of\u0026nbsp;\u0026sim;56.6 electrons \u0026Aring;\u003csup\u003e-2\u003c/sup\u003e at the sample. Using a defocus range of -1.0 to -2.5 \u0026mu;m, 3306 movie stacks were collected.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eImage processing\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eFor the EmrAB-TolC pump-EA, the super-resolution movie frames were adjusted, including correction for gain reference and 2\u0026acute;\u0026nbsp;binning, followed by motion correction using MOTIONCORR2 \u003csup\u003e49\u003c/sup\u003e. Merging of the aligned movie frames into micrographs then allowed estimation of the contrast transfer function (CTF) using CTFFIND4 \u003csup\u003e50\u003c/sup\u003e. RELION v3.1.3 software was used for subsequent image processing steps\u003csup\u003e51,52\u003c/sup\u003e. Templates for automatic particle picking were derived from (reference-free) two-dimensional classification of a manually selected particle subset. To minimize reference bias, these templates underwent low-pass filtering to 20 \u0026Aring;, enabling the automatic selection of 816,894 particles from all the micrographs. Two-dimensional classification of the images yielded 480,902 suitable particles.\u003c/p\u003e\n\u003cp\u003eAdditional image processing steps were carried out using CryoSPARC \u003csup\u003e53\u003c/sup\u003e. Particles chosen via two-dimensional classification within RELION were imported into CryoSPARC and subjected to two further rounds of two-dimensional classification to discard any obviously discrepant particles, yielding 56,006 particles for classification by \u003cem\u003eab initio\u003c/em\u003e reconstruction using CryoSPARC. The parameters for this classification included two-class \u003cem\u003eab initio\u003c/em\u003e reconstruction with specific settings (initial alignment resolution of 25 \u0026Aring;; maximum alignment resolution of 6 \u0026Aring;; initial minibatch size of 150; final size of 600; and class similarity of 0) \u003csup\u003e54\u003c/sup\u003e. A total of 33,460 particles contributed to the resulting 3D volume, which was subjected to nonuniform refinement, generating an EmrAB-TolC reconstruction at a resolution of 3.14 \u0026Aring;. The resulting map quality for the EmrA and TolC sections was high, but the quality for the EmrB section was low. The AlphaFold2 \u003csup\u003e55\u003c/sup\u003e-generated EmrB model was subsequently transformed into a map using e2pdb2mrc.py in EMAN2 \u003csup\u003e56\u003c/sup\u003e and subsequently aligned to the EmrB map section within the EmrAB-TolC reconstruction. A focused mask on EmrB was applied to carry out 3D classification (without alignment), yielding five quality classes from 10. Aligning the 3D volumes and corresponding 19,180 particles from these classes using the Align 3D Maps program in CryoSPARC resulted in a homogeneous reconstruction of EmrAB-TolC at a resolution of 3.45 angstroms, with an improved map quality for the EmrB section. Local refinement was then carried out using a focused mask on the EmrAB portion, leading to a refined EmrAB-TolC reconstruction at 3.13 \u0026Aring; with improved map quality for the EmrAB sections (map-1 of pump-EA; and Extended Data Fig. 2c, d). Local refinement was also conducted using a focused mask on the TolC and \u0026alpha;-helical hairpin domains of EmrA, leading to a refined reconstruction at 3.14 \u0026Aring; with better map quality for the EmrA-TolC sections (map-2 of pump-EA; and Extended Data Fig. 2c, d).\u003c/p\u003e\n\u003cp\u003eA similar image processing procedure was employed for EmrAB-TolC pump-FA, which generated a reconstruction at 3.59 \u0026Aring; resolution (Extended Data Fig. 3c, d; Extended Data Table 2).\u003c/p\u003e\n\u003cp\u003eResolution estimation was accomplished via CryoSPARC using independently refined half-reconstructions. The criterion for Fourier shell correlation was set at 0.143. The ResMap wrapper in CryoSPARC was used to calculate variation in local resolution (Extended Data Fig. \u0026nbsp;2; Extended Data Fig. 3) \u003csup\u003e57\u003c/sup\u003e.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eProtein structure prediction and modelling\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAlphaFold3, AlphaFold2 and AF-Cluster were all employed for protein structure predictions. High-confidence predictions were selected based on their per-residue confidence score and corresponding predicted aligned error plot. Unless stated otherwise, the top rank of 5 predictions is shown and is visualized in ChimeraX Version 1.7. AlphaFold3 Beta was used to generate models of full-length EmrB (UniProt P0AEJ0) \u003csup\u003e36\u003c/sup\u003e. AlphaFold3-Multimer Beta was used to generate models of EmrAB, consisting of one molecule of full-length EmrB (UniProt P0AEJ0) with six molecules of full-length EmrA (UniProt P27303). AF-Cluster was used to predict multiple conformations of EmrB \u003csup\u003e37\u003c/sup\u003e. An MSA of EmrB (UniProt P0AEJ0) consisting of ~6000 sequences was first generated using ColabFold \u003csup\u003e58\u003c/sup\u003e. These sequences were then clustered into ~300 groups based on similarity using DBSCAN. A total of 16 clusters of sufficient size (at least 30 sequences) were then subsequently predicted using AlphaFold2. These conformations were visualized and analyzed using both ChimeraX and ChimeraUCSF.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eModel building and refinement\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe two half-maps of cryo-EM map-1 for EmrAB-TolC pump-EA were used to perform local map sharpening with DeepEMhancer \u003csup\u003e59\u003c/sup\u003e. The resulting full map was segmented into three TolC protomers, six EmrA protomers, and one EmrB protomer. A model of the \u0026beta;-barrel domain without \u0026beta;-CL, and the lipoyl domain, generated by AlphaFold3 \u003csup\u003e36\u003c/sup\u003e, was fitted to the map of individual EmrA protomers using Chimera \u003csup\u003e60\u003c/sup\u003e. The \u0026alpha;-helical hairpin domain, C-terminal \u0026alpha;-helix, and \u0026beta;-CL of the \u0026beta;-barrel domain were manually built using Coot \u003csup\u003e61\u003c/sup\u003e. Each EmrA protomer model was subsequently refined with Rosetta \u003csup\u003e62\u003c/sup\u003e.\u003c/p\u003e\n\u003cp\u003eA homology model of EmrB in the outward-open state was generated by AlphaFold3 \u003csup\u003e36\u003c/sup\u003e. Chimera was used to fit the EBN and EBC domains of this homology model to the map.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eThe EmrB model was subsequently refined using Rosetta.\u003c/p\u003e\n\u003cp\u003eThe two half-maps of cryo-EM map-2 for EmrAB-TolC pump-EA were used to conduct local map sharpening using DeepEMhancer. The structure of trimeric TolC, derived from the MacAB-TolC pump (PDB code: 5NIK), was docked into the TolC section of the modified map-2 using Chimera. To improve the local fit to the map, manual adjustments were made to the model using Coot.\u003c/p\u003e\n\u003cp\u003eModels of individual components were fitted into the cryo-EM map-1 of EmrAB-TolC using Chimera. Adjustments were made to the entire EmrAB-TolC pump-EA model to reduce the clashes using ISOLDE \u003csup\u003e63\u003c/sup\u003e. The model was then undergone real-space refinement against this modified map in the Phenix package \u003csup\u003e64\u003c/sup\u003e. The cryo-EM map-1 of EmrAB-TolC pump-EA without global B-factor sharpening was then used to perform model-based local map sharpening with LocScale \u003csup\u003e65\u003c/sup\u003e.\u003c/p\u003e\n\u003cp\u003eThe two half-maps of cryo-EM map for EmrAB-TolC pump-FA were used to perform local map sharpening with DeepEMhancer. The refined model of EmrAB-TolC pump-EA was fitted into the cryo-EM map of pump-FA. The model section of the N-terminal TM helix of EmrA, generated by AlphaFold3, was fitted into the map of the individual EmrA protomer. Model sections without defined map were deleted. To improve the local fit to the map, manual adjustments were made to the model using Coot. The cryo-EM map of EmrAB-TolC pump-FA without global B-factor sharpening was then used to perform model-based local map sharpening with LocScale.\u003c/p\u003e\n\u003cp\u003eAny Ramachandran outliers were manually corrected in Coot, and stereochemistry was ensured using MolProbity (Extended Data Table 2)\u003csup\u003e66\u003c/sup\u003e.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eData availability\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eCoordinates have been deposited in the Protein Data Bank (PDB) under PDB codes 8ZAL (EmrAB-TolC pump-EA) and 8ZAR (EmrAB-TolC pump-FA). Cryo-EM maps have been deposited in the Electron Microscopy Data Bank (EMD) under EMD codes EMD-39879 (EmrAB-TolC pump-EA) and EMD-39885 (EmrAB-TolC pump-FA).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAcknowledgements:\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis work was supported by the National Key R\u0026amp;D Program of China (2022YFC2303200); National Natural Science Foundation of China (31971133 to D.D.; 32270064 and 92478118 to Y.C.), the Science and Technology Commission of Shanghai Municipality (19PJ1407900, 19JC1414000 and 22WZ2504100 to D.D.; 24ZR1493200 to Y.C.), and the Chinese Academy of Sciences (XDB0570000 and 176002GJHZ2022022MI to Y.C.). BFL was supported by ERC Advanced grant (742210) and a Wellcome Trust Investigator award (200873/Z/16/Z). Cryo-EM data were collected at the Bio-Electron Microscopy Facility of ShanghaiTech University with the assistance of Q. Sun, D. Liu, Z. Zhang, L. Wang and Y. Yang. We thank the Molecular Imaging Core Facility, the Molecular and Cell Biology Core Facility, and the Multi-Omics Core Facility at the School of Life Science and Technology for providing technical support. We are also grateful for the support of Lajos Kalmar of the MRC Toxicology Unit in the use of high-performance computing used in this study.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthor contributions:\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eZ.Z. performed cloning and overexpression of the EmrAB-TolC complex; Z.Z. and T.M. purified the EmrAB-TolC complex, prepared cryo-EM samples, collected cryo-EM data, determined structures, performed model building and refinement and prepared figures for the manuscript; X.G., W.S., H.Z., J.G., S.L., H.L. and Q.O. optimised in-column peptide-disc methods, prepared homemade graphene monolayer grids, and assisted the collection of cryo-EM data; R.D., H.J., Z.Z., T.M., \u0026nbsp;X.G., S.Z. and W.S. performed antibiotics sensitivity assays; J.G., S.L., Z.Z. and T.M. performed drug-proton antiport assay; T.M. and H.L. carried out Western blot assay; Y.C. supervised antibiotics sensitivity assays and discussed project design; M.L.J. performed protein structure predictions and modelling;D.D. and B.L. conceived the project; D.D. designed and supervised all experiments; D.D. and B.L. wrote the manuscript. All the authors contributed to the data interpretation and manuscript preparation.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eDeclaration of interests:\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors declare no competing interests.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eHodges, F. J., Torres, V. V. L., Cunningham, A. F., Henderson, I. R. \u0026amp; Icke, C. Redefining the bacterial Type I protein secretion system. \u003cem\u003eAdv Microb Physiol\u003c/em\u003e \u003cstrong\u003e82\u003c/strong\u003e, 155-204 (2023). https://doi.org/10.1016/bs.ampbs.2022.10.003\u003c/li\u003e\n\u003cli\u003eCosta, T. R.\u003cem\u003e et al.\u003c/em\u003e Secretion systems in Gram-negative bacteria: structural and mechanistic insights. \u003cem\u003eNat Rev Microbiol\u003c/em\u003e \u003cstrong\u003e13\u003c/strong\u003e, 343-359 (2015). https://doi.org/10.1038/nrmicro3456\u003c/li\u003e\n\u003cli\u003eDu, D., van Veen, H. W. \u0026amp; Luisi, B. F. Assembly and operation of bacterial tripartite multidrug efflux pumps. \u003cem\u003eTrends Microbiol\u003c/em\u003e \u003cstrong\u003e23\u003c/strong\u003e, 311-319 (2015). https://doi.org/10.1016/j.tim.2015.01.010\u003c/li\u003e\n\u003cli\u003eAlav, I.\u003cem\u003e et al.\u003c/em\u003e Structure, Assembly, and Function of Tripartite Efflux and Type 1 Secretion Systems in Gram-Negative Bacteria. \u003cem\u003eChem Rev\u003c/em\u003e \u003cstrong\u003e121\u003c/strong\u003e, 5479-5596 (2021). https://doi.org/10.1021/acs.chemrev.1c00055\u003c/li\u003e\n\u003cli\u003eWang, Z.\u003cem\u003e et al.\u003c/em\u003e An allosteric transport mechanism for the AcrAB-TolC multidrug efflux pump. \u003cem\u003eElife\u003c/em\u003e \u003cstrong\u003e6\u003c/strong\u003e (2017). https://doi.org/10.7554/eLife.24905\u003c/li\u003e\n\u003cli\u003eDu, D.\u003cem\u003e et al.\u003c/em\u003e Structure of the AcrAB-TolC multidrug efflux pump. \u003cem\u003eNature\u003c/em\u003e \u003cstrong\u003e509\u003c/strong\u003e, 512-515 (2014). https://doi.org/10.1038/nature13205\u003c/li\u003e\n\u003cli\u003eFitzpatrick, A. W. P.\u003cem\u003e et al.\u003c/em\u003e Structure of the MacAB-TolC ABC-type tripartite multidrug efflux pump. \u003cem\u003eNat Microbiol\u003c/em\u003e \u003cstrong\u003e2\u003c/strong\u003e, 17070 (2017). https://doi.org/10.1038/nmicrobiol.2017.70\u003c/li\u003e\n\u003cli\u003eDu, D.\u003cem\u003e et al.\u003c/em\u003e Multidrug efflux pumps: structure, function and regulation. \u003cem\u003eNat Rev Microbiol\u003c/em\u003e \u003cstrong\u003e16\u003c/strong\u003e, 523-539 (2018). https://doi.org/10.1038/s41579-018-0048-6\u003c/li\u003e\n\u003cli\u003eYousefian, N.\u003cem\u003e et al.\u003c/em\u003e Structural characterization of the EmrAB-TolC efflux complex from E. coli. \u003cem\u003eBiochim Biophys Acta Biomembr\u003c/em\u003e \u003cstrong\u003e1863\u003c/strong\u003e, 183488 (2021). https://doi.org/10.1016/j.bbamem.2020.183488\u003c/li\u003e\n\u003cli\u003eVarela, M. F., Stephen, J., Bharti, D., Lekshmi, M. \u0026amp; Kumar, S. Inhibition of Multidrug Efflux Pumps Belonging to the Major Facilitator Superfamily in Bacterial Pathogens. \u003cem\u003eBiomedicines\u003c/em\u003e \u003cstrong\u003e11\u003c/strong\u003e (2023). https://doi.org/10.3390/biomedicines11051448\u003c/li\u003e\n\u003cli\u003eLin, M. F., Lin, Y. Y. \u0026amp; Lan, C. Y. Contribution of EmrAB efflux pumps to colistin resistance in Acinetobacter baumannii. \u003cem\u003eJ Microbiol\u003c/em\u003e \u003cstrong\u003e55\u003c/strong\u003e, 130-136 (2017). https://doi.org/10.1007/s12275-017-6408-5\u003c/li\u003e\n\u003cli\u003ePuertolas-Balint, F., Warsi, O., Linkevicius, M., Tang, P. C. \u0026amp; Andersson, D. I. Mutations that increase expression of the EmrAB-TolC efflux pump confer increased resistance to nitroxoline in Escherichia coli. \u003cem\u003eJ Antimicrob Chemother\u003c/em\u003e \u003cstrong\u003e75\u003c/strong\u003e, 300-308 (2020). https://doi.org/10.1093/jac/dkz434\u003c/li\u003e\n\u003cli\u003eLomovskaya, O. L., K. Emr, an Escherichia coli locus for multidrug resistance. \u003cem\u003eCell Biology\u003c/em\u003e \u003cstrong\u003e89\u003c/strong\u003e (1992).\u003c/li\u003e\n\u003cli\u003eHinchliffe, P.\u003cem\u003e et al.\u003c/em\u003e Structure of the periplasmic adaptor protein from a major facilitator superfamily (MFS) multidrug efflux pump. \u003cem\u003eFEBS Lett\u003c/em\u003e \u003cstrong\u003e588\u003c/strong\u003e, 3147-3153 (2014). https://doi.org/10.1016/j.febslet.2014.06.055\u003c/li\u003e\n\u003cli\u003eJiang, D.\u003cem\u003e et al.\u003c/em\u003e Structure of the YajR transporter suggests a transport mechanism based on the conserved motif A. \u003cem\u003eProc Natl Acad Sci U S A\u003c/em\u003e \u003cstrong\u003e110\u003c/strong\u003e, 14664-14669 (2013). https://doi.org/10.1073/pnas.1308127110\u003c/li\u003e\n\u003cli\u003eYin, Y., He, X., Szewczyk, P., Nguyen, T. \u0026amp; Chang, G. Structure of the Multidrug Transporter EmrD from \u003cem\u003eEscherichia coli\u003c/em\u003e. \u003cem\u003eScience\u003c/em\u003e \u003cstrong\u003e312\u003c/strong\u003e, 741-744 (2006). https://doi.org/10.1126/science.1125629\u003c/li\u003e\n\u003cli\u003eNagarathinam, K.\u003cem\u003e et al.\u003c/em\u003e Outward open conformation of a Major Facilitator Superfamily multidrug/H+ antiporter provides insights into switching mechanism. \u003cem\u003eNature Communications\u003c/em\u003e \u003cstrong\u003e9\u003c/strong\u003e (2018). https://doi.org/10.1038/s41467-018-06306-x\u003c/li\u003e\n\u003cli\u003eHeng, J.\u003cem\u003e et al.\u003c/em\u003e Substrate-bound structure of the E. coli multidrug resistance transporter MdfA. \u003cem\u003eCell Research\u003c/em\u003e \u003cstrong\u003e25\u003c/strong\u003e, 1060-1073 (2015). https://doi.org/10.1038/cr.2015.94\u003c/li\u003e\n\u003cli\u003eMajumder, P.\u003cem\u003e et al.\u003c/em\u003e Dissection of Protonation Sites for Antibacterial Recognition and Transport in QacA, a Multi-Drug Efflux Transporter. \u003cem\u003eJ Mol Biol\u003c/em\u003e \u003cstrong\u003e431\u003c/strong\u003e, 2163-2179 (2019). https://doi.org/10.1016/j.jmb.2019.03.015\u003c/li\u003e\n\u003cli\u003eRemm, S.\u003cem\u003e et al.\u003c/em\u003e Structural basis for triacylglyceride extraction from mycobacterial inner membrane by MFS transporter Rv1410. \u003cem\u003eNature Communications\u003c/em\u003e \u003cstrong\u003e14\u003c/strong\u003e (2023). https://doi.org/10.1038/s41467-023-42073-0\u003c/li\u003e\n\u003cli\u003eTanabe, M.\u003cem\u003e et al.\u003c/em\u003e The multidrug resistance efflux complex, EmrAB from Escherichia coli forms a dimer in vitro. \u003cem\u003eBiochem Biophys Res Commun\u003c/em\u003e \u003cstrong\u003e380\u003c/strong\u003e, 338-342 (2009). https://doi.org/10.1016/j.bbrc.2009.01.081\u003c/li\u003e\n\u003cli\u003eSymmons, M. F., Marshall, R. L. \u0026amp; Bavro, V. N. Architecture and roles of periplasmic adaptor proteins in tripartite efflux assemblies. \u003cem\u003eFrontiers in Microbiology\u003c/em\u003e \u003cstrong\u003e6\u003c/strong\u003e (2015). https://doi.org/10.3389/fmicb.2015.00513\u003c/li\u003e\n\u003cli\u003eChun, E.\u003cem\u003e et al.\u003c/em\u003e Fusion partner toolchest for the stabilization and crystallization of G protein-coupled receptors. \u003cem\u003eStructure\u003c/em\u003e \u003cstrong\u003e20\u003c/strong\u003e, 967-976 (2012). https://doi.org/10.1016/j.str.2012.04.010\u003c/li\u003e\n\u003cli\u003eGotzke, H.\u003cem\u003e et al.\u003c/em\u003e The ALFA-tag is a highly versatile tool for nanobody-based bioscience applications. \u003cem\u003eNat Commun\u003c/em\u003e \u003cstrong\u003e10\u003c/strong\u003e, 4403 (2019). https://doi.org/10.1038/s41467-019-12301-7\u003c/li\u003e\n\u003cli\u003eXie, P.\u003cem\u003e et al.\u003c/em\u003e A fiducial-assisted strategy compatible with resolving small MFS transporter structures in multiple conformations using cryo-EM. \u003cem\u003eNat Commun\u003c/em\u003e \u003cstrong\u003e16\u003c/strong\u003e, 7 (2025). https://doi.org/10.1038/s41467-024-54986-5\u003c/li\u003e\n\u003cli\u003eHiggins, M. K., Bokma, E., Koronakis, E., Hughes, C. \u0026amp; Koronakis, V. Structure of the periplasmic component of a bacterial drug efflux pump. \u003cem\u003eProc Natl Acad Sci U S A\u003c/em\u003e \u003cstrong\u003e101\u003c/strong\u003e, 9994-9999 (2004). https://doi.org/10.1073/pnas.0400375101\u003c/li\u003e\n\u003cli\u003eDaury, L.\u003cem\u003e et al.\u003c/em\u003e Tripartite assembly of RND multidrug efflux pumps. \u003cem\u003eNat Commun\u003c/em\u003e \u003cstrong\u003e7\u003c/strong\u003e, 10731 (2016). https://doi.org/10.1038/ncomms10731\u003c/li\u003e\n\u003cli\u003eSulavik, M. C.\u003cem\u003e et al.\u003c/em\u003e Antibiotic susceptibility profiles of Escherichia coli strains lacking multidrug efflux pump genes. \u003cem\u003eAntimicrob Agents Chemother\u003c/em\u003e \u003cstrong\u003e45\u003c/strong\u003e, 1126-1136 (2001). https://doi.org/10.1128/AAC.45.4.1126-1136.2001\u003c/li\u003e\n\u003cli\u003eSu, C. C.\u003cem\u003e et al.\u003c/em\u003e Crystal structure of the CusBA heavy-metal efflux complex of Escherichia coli. \u003cem\u003eNature\u003c/em\u003e \u003cstrong\u003e470\u003c/strong\u003e, 558-562 (2011). https://doi.org/10.1038/nature09743\u003c/li\u003e\n\u003cli\u003eZhao, Y.\u003cem\u003e et al.\u003c/em\u003e Crystal Structure of the E. coli Peptide Transporter YbgH. \u003cem\u003eStructure\u003c/em\u003e \u003cstrong\u003e22\u003c/strong\u003e, 1152-1160 (2014). https://doi.org/10.1016/j.str.2014.06.008\u003c/li\u003e\n\u003cli\u003eNewstead, S.\u003cem\u003e et al.\u003c/em\u003e Crystal structure of a prokaryotic homologue of the mammalian oligopeptide-proton symporters, PepT1 and PepT2. \u003cem\u003eEMBO J\u003c/em\u003e \u003cstrong\u003e30\u003c/strong\u003e, 417-426 (2011). https://doi.org/10.1038/emboj.2010.309\u003c/li\u003e\n\u003cli\u003eFowler, P. W.\u003cem\u003e et al.\u003c/em\u003e Gating topology of the proton-coupled oligopeptide symporters. \u003cem\u003eStructure\u003c/em\u003e \u003cstrong\u003e23\u003c/strong\u003e, 290-301 (2015). https://doi.org/10.1016/j.str.2014.12.012\u003c/li\u003e\n\u003cli\u003eJin, J., Guffanti, A. A., Beck, C. \u0026amp; Krulwich, T. A. Twelve-Transmembrane-Segment (TMS) Version (\u0026Delta;TMS VII-VIII) of the 14-TMS Tet(L) Antibiotic Resistance Protein Retains Monovalent Cation Transport Modes but Lacks Tetracycline Efflux Capacity. \u003cem\u003eJournal of Bacteriology\u003c/em\u003e \u003cstrong\u003e183\u003c/strong\u003e, 2667-2671 (2001). https://doi.org/10.1128/jb.183.8.2667-2671.2001\u003c/li\u003e\n\u003cli\u003eHenderson, P. J. F.\u003cem\u003e et al.\u003c/em\u003e Physiological Functions of Bacterial \u0026quot;Multidrug\u0026quot; Efflux Pumps. \u003cem\u003eChem Rev\u003c/em\u003e \u003cstrong\u003e121\u003c/strong\u003e, 5417-5478 (2021). https://doi.org/10.1021/acs.chemrev.0c01226\u003c/li\u003e\n\u003cli\u003eDrew, D., North, R. A., Nagarathinam, K. \u0026amp; Tanabe, M. Structures and General Transport Mechanisms by the Major Facilitator Superfamily (MFS). \u003cem\u003eChem Rev\u003c/em\u003e \u003cstrong\u003e121\u003c/strong\u003e, 5289-5335 (2021). https://doi.org/10.1021/acs.chemrev.0c00983\u003c/li\u003e\n\u003cli\u003eAbramson, J.\u003cem\u003e et al.\u003c/em\u003e Accurate structure prediction of biomolecular interactions with AlphaFold 3. \u003cem\u003eNature\u003c/em\u003e (2024). https://doi.org/10.1038/s41586-024-07487-w\u003c/li\u003e\n\u003cli\u003eWayment-Steele, H. K.\u003cem\u003e et al.\u003c/em\u003e Predicting multiple conformations via sequence clustering and AlphaFold2. \u003cem\u003eNature\u003c/em\u003e \u003cstrong\u003e625\u003c/strong\u003e, 832-839 (2024). https://doi.org/10.1038/s41586-023-06832-9\u003c/li\u003e\n\u003cli\u003eSauve, S., Williamson, J., Polasa, A. \u0026amp; Moradi, M. Ins and Outs of Rocker Switch Mechanism in Major Facilitator Superfamily of Transporters. \u003cem\u003eMembranes (Basel)\u003c/em\u003e \u003cstrong\u003e13\u003c/strong\u003e (2023). https://doi.org/10.3390/membranes13050462\u003c/li\u003e\n\u003cli\u003eGlavier, M.\u003cem\u003e et al.\u003c/em\u003e Antibiotic export by MexB multidrug efflux transporter is allosterically controlled by a MexA-OprM chaperone-like complex. \u003cem\u003eNat Commun\u003c/em\u003e \u003cstrong\u003e11\u003c/strong\u003e, 4948 (2020). https://doi.org/10.1038/s41467-020-18770-5\u003c/li\u003e\n\u003cli\u003eWu, H.-H., Symersky, J. \u0026amp; Lu, M. Structure and mechanism of a redesigned multidrug transporter from the Major Facilitator Superfamily. \u003cem\u003eScientific Reports\u003c/em\u003e \u003cstrong\u003e10\u003c/strong\u003e (2020). https://doi.org/10.1038/s41598-020-60332-8\u003c/li\u003e\n\u003cli\u003eKim, J.-S.\u003cem\u003e et al.\u003c/em\u003e Crystal Structure of a Soluble Fragment of the Membrane Fusion Protein HlyD in a Type I Secretion System of Gram-Negative Bacteria. \u003cem\u003eStructure\u003c/em\u003e \u003cstrong\u003e24\u003c/strong\u003e, 477-485 (2016). https://doi.org/10.1016/j.str.2015.12.012\u003c/li\u003e\n\u003cli\u003eTeelucksingh, T.\u003cem\u003e et al.\u003c/em\u003e A genetic platform to investigate the functions of bacterial drug efflux pumps. \u003cem\u003eNat Chem Biol\u003c/em\u003e \u003cstrong\u003e18\u003c/strong\u003e, 1399-1409 (2022). https://doi.org/10.1038/s41589-022-01119-y\u003c/li\u003e\n\u003cli\u003eAngiulli, G.\u003cem\u003e et al.\u003c/em\u003e New approach for membrane protein reconstitution into peptidiscs and basis for their adaptability to different proteins. \u003cem\u003eElife\u003c/em\u003e \u003cstrong\u003e9\u003c/strong\u003e (2020). https://doi.org/10.7554/eLife.53530\u003c/li\u003e\n\u003cli\u003eCarlson, M. L.\u003cem\u003e et al.\u003c/em\u003e The Peptidisc, a simple method for stabilizing membrane proteins in detergent-free solution. \u003cem\u003eElife\u003c/em\u003e \u003cstrong\u003e7\u003c/strong\u003e (2018). https://doi.org/10.7554/eLife.34085\u003c/li\u003e\n\u003cli\u003eFluman, N., Ryan, C. M., Whitelegge, J. P. \u0026amp; Bibi, E. Dissection of mechanistic principles of a secondary multidrug efflux protein. \u003cem\u003eMol Cell\u003c/em\u003e \u003cstrong\u003e47\u003c/strong\u003e, 777-787 (2012). https://doi.org/10.1016/j.molcel.2012.06.018\u003c/li\u003e\n\u003cli\u003eBogdanov, M. Preparation of Uniformly Oriented Inverted Inner (Cytoplasmic) Membrane Vesicles from Gram-Negative Bacterial Cells. \u003cem\u003eMethods Mol Biol\u003c/em\u003e \u003cstrong\u003e2715\u003c/strong\u003e, 159-180 (2024). https://doi.org/10.1007/978-1-0716-3445-5_10\u003c/li\u003e\n\u003cli\u003eVerkhovskaya, M. Preparation of Everted Membrane Vesicles from Escherichia coli Cells. \u003cem\u003eBio Protoc\u003c/em\u003e \u003cstrong\u003e7\u003c/strong\u003e, e2254 (2017). https://doi.org/10.21769/BioProtoc.2254\u003c/li\u003e\n\u003cli\u003eSchorb, M., Haberbosch, I., Hagen, W. J. H., Schwab, Y. \u0026amp; Mastronarde, D. N. Software tools for automated transmission electron microscopy. \u003cem\u003eNature Methods\u003c/em\u003e \u003cstrong\u003e16\u003c/strong\u003e, 471-477 (2019). https://doi.org/10.1038/s41592-019-0396-9\u003c/li\u003e\n\u003cli\u003eLi, X.\u003cem\u003e et al.\u003c/em\u003e Electron counting and beam-induced motion correction enable near-atomic-resolution single-particle cryo-EM. \u003cem\u003eNature Methods\u003c/em\u003e \u003cstrong\u003e10\u003c/strong\u003e, 584-590 (2013). https://doi.org/10.1038/nmeth.2472\u003c/li\u003e\n\u003cli\u003eRohou, A. \u0026amp; Grigorieff, N. CTFFIND4: Fast and accurate defocus estimation from electron micrographs. \u003cem\u003eJournal of Structural Biology\u003c/em\u003e \u003cstrong\u003e192\u003c/strong\u003e, 216-221 (2015). https://doi.org/10.1016/j.jsb.2015.08.008\u003c/li\u003e\n\u003cli\u003eScheres, S. H. RELION: implementation of a Bayesian approach to cryo-EM structure determination. \u003cem\u003eJ Struct Biol\u003c/em\u003e \u003cstrong\u003e180\u003c/strong\u003e, 519-530 (2012). https://doi.org/10.1016/j.jsb.2012.09.006\u003c/li\u003e\n\u003cli\u003eZivanov, J.\u003cem\u003e et al.\u003c/em\u003e New tools for automated high-resolution cryo-EM structure determination in RELION-3. \u003cem\u003eElife\u003c/em\u003e \u003cstrong\u003e7\u003c/strong\u003e (2018). https://doi.org/10.7554/eLife.42166\u003c/li\u003e\n\u003cli\u003ePunjani, A., Rubinstein, J. L., Fleet, D. J. \u0026amp; Brubaker, M. A. cryoSPARC: algorithms for rapid unsupervised cryo-EM structure determination. \u003cem\u003eNat Methods\u003c/em\u003e \u003cstrong\u003e14\u003c/strong\u003e, 290-296 (2017). https://doi.org/10.1038/nmeth.4169\u003c/li\u003e\n\u003cli\u003eWright, N. J.\u003cem\u003e et al.\u003c/em\u003e Methotrexate recognition by the human reduced folate carrier SLC19A1. \u003cem\u003eNature\u003c/em\u003e \u003cstrong\u003e609\u003c/strong\u003e, 1056-1062 (2022). https://doi.org/10.1038/s41586-022-05168-0\u003c/li\u003e\n\u003cli\u003eJumper, J.\u003cem\u003e et al.\u003c/em\u003e Highly accurate protein structure prediction with AlphaFold. \u003cem\u003eNature\u003c/em\u003e \u003cstrong\u003e596\u003c/strong\u003e, 583-589 (2021). https://doi.org/10.1038/s41586-021-03819-2\u003c/li\u003e\n\u003cli\u003eTang, G.\u003cem\u003e et al.\u003c/em\u003e EMAN2: an extensible image processing suite for electron microscopy. \u003cem\u003eJ Struct Biol\u003c/em\u003e \u003cstrong\u003e157\u003c/strong\u003e, 38-46 (2007). https://doi.org/10.1016/j.jsb.2006.05.009\u003c/li\u003e\n\u003cli\u003eKucukelbir, A., Sigworth, F. J. \u0026amp; Tagare, H. D. Quantifying the local resolution of cryo-EM density maps. \u003cem\u003eNature Methods\u003c/em\u003e \u003cstrong\u003e11\u003c/strong\u003e, 63-65 (2014). https://doi.org/10.1038/nmeth.2727\u003c/li\u003e\n\u003cli\u003eMirdita, M.\u003cem\u003e et al.\u003c/em\u003e ColabFold: making protein folding accessible to all. \u003cem\u003eNature Methods\u003c/em\u003e \u003cstrong\u003e19\u003c/strong\u003e, 679-682 (2022). https://doi.org/10.1038/s41592-022-01488-1\u003c/li\u003e\n\u003cli\u003eSanchez-Garcia, R.\u003cem\u003e et al.\u003c/em\u003e DeepEMhancer: a deep learning solution for cryo-EM volume post-processing. \u003cem\u003eCommun Biol\u003c/em\u003e \u003cstrong\u003e4\u003c/strong\u003e, 874 (2021). https://doi.org/10.1038/s42003-021-02399-1\u003c/li\u003e\n\u003cli\u003ePettersen, E. F.\u003cem\u003e et al.\u003c/em\u003e UCSF Chimera--a visualization system for exploratory research and analysis. \u003cem\u003eJ Comput Chem\u003c/em\u003e \u003cstrong\u003e25\u003c/strong\u003e, 1605-1612 (2004). https://doi.org/10.1002/jcc.20084\u003c/li\u003e\n\u003cli\u003eEmsley, P., Lohkamp, B., Scott, W. G. \u0026amp; Cowtan, K. Features and development of Coot. \u003cem\u003eActa Crystallogr D Biol Crystallogr\u003c/em\u003e \u003cstrong\u003e66\u003c/strong\u003e, 486-501 (2010). https://doi.org/10.1107/S0907444910007493\u003c/li\u003e\n\u003cli\u003eWang, R. Y.\u003cem\u003e et al.\u003c/em\u003e Automated structure refinement of macromolecular assemblies from cryo-EM maps using Rosetta. \u003cem\u003eElife\u003c/em\u003e \u003cstrong\u003e5\u003c/strong\u003e (2016). https://doi.org/10.7554/eLife.17219\u003c/li\u003e\n\u003cli\u003eCroll, T. I. ISOLDE: a physically realistic environment for model building into low-resolution electron-density maps. \u003cem\u003eActa Crystallogr D Struct Biol\u003c/em\u003e \u003cstrong\u003e74\u003c/strong\u003e, 519-530 (2018). https://doi.org/10.1107/S2059798318002425\u003c/li\u003e\n\u003cli\u003eLiebschner, D.\u003cem\u003e et al.\u003c/em\u003e Macromolecular structure determination using X-rays, neutrons and electrons: recent developments in Phenix. \u003cem\u003eActa Crystallogr D Struct Biol\u003c/em\u003e \u003cstrong\u003e75\u003c/strong\u003e, 861-877 (2019). https://doi.org/10.1107/S2059798319011471\u003c/li\u003e\n\u003cli\u003eJakobi, A. J., Wilmanns, M. \u0026amp; Sachse, C. Model-based local density sharpening of cryo-EM maps. \u003cem\u003eElife\u003c/em\u003e \u003cstrong\u003e6\u003c/strong\u003e (2017). https://doi.org/10.7554/eLife.27131\u003c/li\u003e\n\u003cli\u003eWilliams, C. J.\u003cem\u003e et al.\u003c/em\u003e MolProbity: More and better reference data for improved all-atom structure validation. \u003cem\u003eProtein Sci\u003c/em\u003e \u003cstrong\u003e27\u003c/strong\u003e, 293-315 (2018). https://doi.org/10.1002/pro.3330\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":true,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"nature-portfolio","isNatureJournal":true,"hasQc":false,"allowDirectSubmit":false,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"","title":"Nature Portfolio","twitterHandle":"","acdcEnabled":false,"dfaEnabled":false,"editorialSystem":"ejp","reportingPortfolio":"","inReviewEnabled":true,"inReviewRevisionsEnabled":false},"keywords":"","lastPublishedDoi":"10.21203/rs.3.rs-5901545/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-5901545/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eTransport of proteins and small molecules across cellular membrane is crucial for bacterial interaction with the environment and survival against antibiotics. In Gram-negative bacteria that possess two layers of membranes, specialized macromolecular machines are required to transport substrates across the envelope, often via an indirect stepwise procedure. The major facilitator superfamily (MFS)-type tripartite efflux pumps and type I secretion systems likely employ a direct one-step transport approach, but the structural mechanisms remain elusive. A representative MFS-type tripartite efflux pump, EmrAB-TolC, mediates resistance to multiple antimicrobial drugs through proton-coupled EmrB, a member of the DHA2 transporter family. Here, we report the high-resolution (3.13 \u0026Aring;) structure of the EmrAB-TolC pump, revealing a unique, asymmetric architecture emerging from the assembly of TolC:EmrA:EmrB with a ratio of 3:6:1. This structure reveals two critical subdomains AssA and AssB that are essential for the pump assembly. Key residues involved in drug recognition, proton translocation and coupling are identified and corroborated by mutagenesis and antibiotic sensitivity assays. The delineation of the complete translocation pathway unravels the one-step drug transport mechanism of the MFS pump across the entire envelope of Gram-negative bacteria.\u003c/p\u003e","manuscriptTitle":"One-step drug transport across two membranes of Gram-negative bacteria by an MFS tripartite assembly","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-02-13 14:41:44","doi":"10.21203/rs.3.rs-5901545/v1","editorialEvents":[],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"nature-communications","isNatureJournal":true,"hasQc":false,"allowDirectSubmit":false,"externalIdentity":"NCOMMS","sideBox":"Learn more about [Nature Communications](http://www.nature.com/ncomms/)","snPcode":"","submissionUrl":"https://mts-ncomms.nature.com/","title":"Nature Communications","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"ejp","reportingPortfolio":"Nature Communications","inReviewEnabled":true,"inReviewRevisionsEnabled":false}}],"origin":"","ownerIdentity":"5e2e763a-972b-4252-97ba-f3ca5fbc8a22","owner":[],"postedDate":"February 13th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"published-in-journal","subjectAreas":[{"id":44226784,"name":"Biological sciences/Structural biology/Electron microscopy/Cryoelectron microscopy"},{"id":44226785,"name":"Biological sciences/Microbiology/Antimicrobials/Antibiotics"}],"tags":[],"updatedAt":"2026-05-05T07:05:35+00:00","versionOfRecord":{"articleIdentity":"rs-5901545","link":"https://doi.org/10.1038/s41467-026-70500-5","journal":{"identity":"nature-communications","isVorOnly":false,"title":"Nature Communications"},"publishedOn":"2026-03-16 04:00:00","publishedOnDateReadable":"March 16th, 2026"},"versionCreatedAt":"2025-02-13 14:41:44","video":"","vorDoi":"10.1038/s41467-026-70500-5","vorDoiUrl":"https://doi.org/10.1038/s41467-026-70500-5","workflowStages":[]},"version":"v1","identity":"rs-5901545","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-5901545","identity":"rs-5901545","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}