Structural and functional diversity of Resistance-Nodulation-Division (RND) efflux pump transporters with implications for antimicrobial resistance

Abstract

SUMMARYThe discovery of bacterial efflux pumps significantly advanced our understanding of how bacteria can resist cytotoxic compounds that they encounter. Within the structurally and functionally distinct families of efflux pumps, those of the Resistance-Nodulation-Division (RND) superfamily are noteworthy for their ability to reduce the intracellular concentration of structurally diverse antimicrobials. RND systems are possessed by many Gram-negative bacteria, including those causing serious human disease, and frequently contribute to resistance to multiple antibiotics. Herein, we review the current literature on the structure-function relationships of representative transporter proteins of tripartite RND efflux pumps of clinically important pathogens. We emphasize their contribution to bacterial resistance to clinically used antibiotics, host defense antimicrobials and other biocides, as well as highlighting structural similarities and differences among efflux transporters that help bacteria survive in the face of antimicrobials. Furthermore, we discuss technical advances that have facilitated and advanced efflux pump research and suggest future areas of investigation that will advance antimicrobial development efforts.

Keywords: Resistance-Nodulation-Division; antimicrobials; clinical; efflux; resistance; structure-function; transporters.

Fig 1

Prototypical structure of the RND complex and IM transporter. (A) Cryogenic electron microscopy(cryo-EM) structure of the E. coli AcrAB-TolC (PDB 5O66) tripartite complex containing the OM protein TolC, periplasmic adaptor protein AcrA, and IM transporter AcrB. The periplasmic peptidoglycan layer (PG) is also indicated. (B) Zoomed in view of trimeric AcrB highlighting the three conformational states: loose (L; blue), tight (T; purple), and open (O; yellow). In the T conformation, the docking (DC, DN; dark purple), pore (PN1, PN2, PC1, PC2; orange), and the transmembrane (TM; TM1–12; blue) domains are shown. (C) Bottom-up view of panel (B) to show the arrangement of helices TM1–12 in the L/T/O conformation. (D) Rotation of panel (B) to show conserved structural features such as the switch loop (lime green), flexible loop (aqua), and cleft (located between PC1 and PC2). (E) Top-down view of panel (D) with the widest opening of the cleft between PC1 and PC2 (orange) in the L/T/O conformations indicated by dotted line.

Fig 2

Mechanistic model of substrate translocation and proton relay. (A) General schematic of the peristaltic movements of the proximal binding pocket (PBP)/distal binding pocket (DBP) as substrate (orange) moves from the cleft (Ch.2) entrance, through the central cavity passing through the switch loop, to the PAP. In the loose (L) state (blue) substrate moves through the open cleft where the PBP is expanded and exit channel closed. In the tight (T) state (purple), substrate moves from the PBP to an enlarged DBP via a conformational change in the gated switch loop (light green). In the open (O) state (yellow), the cleft is constricted and the exit channel expands, allowing translocation of substrate to the PAP. In the absence of substrate, a resting (R) state (green) has also been observed in which all regions are in a relaxed state. (B) Model of pore (PN1/PN2 and PC1/PC2) and TM (TM4/TM10/TM11) domain changes during proton relay in the L/T/O/R conformational states. In the L state (substrate entry), K940 (TM10) is positioned between deprotonated D408 (TM4) and D407 (TM4). Substrate movement between PN1 and PN2 in the T state is coupled to proton (H⁺) movement in the opposite direction in the TM domain. Proton movement facilitates protonation of D408 resulting in TM4/TM10 repositioning that places K940 in between D407 and R971 (TM11). As substrate is released to the PAP, D408 once again becomes deprotonated and the PMF residue network in the R state resembles that of the L state.

Fig 3

Variations in the TM domain of RND transporters. Differences in the TM region have been observed in two transporter structures: E. coli AcrB (PDB 5O66) and N. gonorrhoeae MtrD (PDB 8DEW). In E. coli, the AcrB RND transporter (white) was found in complex with accessory protein AcrZ (teal). N. gonorrhoeae MtrD was found to have an extended loop positioned between TM9 and TM10 (mapped on PDB:5O66; red) compared to the E. coli TM9/TM10 (blue).

Fig 4

Conformational states observed in RND transporter structures. Transporters are displayed as one-third parts of a circle from a top-down view of the trimer with protomer conformations indicated: L (blue), T (purple), O (yellow), or R (green). (A) E. coli AcrB was found as a symmetric trimer comprising three protomers in L/L/L states and as an asymmetric trimer adopting L/T/O states in the absence of substrates and with bound substrate (orange) structures in the L and/or T state. A symmetric T/T/T state was observed with the inhibitor MBX3132 bound (red). (B) P. aeruginosa MexB was found in L/T/O states both in the absence and presence of substrates, with the latter binding in the T state. (C) A. baumannii AdeB shows high variability in conformational states, adopting a symmetric O/O/O state without substrate and O/O/T, O/R/T, or L/T/O in the presence of substrate. (D) Both B. pseudomallei BpeF and K. pneumoniae OqxB adopt symmetric T/T/T states with substrate bound in all the protomers. (E) N. gonorrhoeae MtrD adopts L/T/O states in the presence of substrates which bind in the T state.

Fig 5

Regions important for RND substrate entry, binding, and specificity. (A) Locations of substrate entry channels in E. coli AcrB (PDB 5O66). Ch.1 (yellow) is located between PC2 and TM8 and moves substrates to the PBP (pink) from the outer leaflet of the IM. Entry to Ch.2/cleft (purple) is located between PC1 and PC2 and moves substrates from the periplasm to the PBP. Ch.3 (green) is located within the central cavity and moves substrates directly to the DBP (gray). Ch.4 (blue) is located between TM1 and TM2 and potentially moves substrates from the outer leaflet of the IM into the DBP. (B) Important regions for substrate recognition within the pore subdomain (orange). Inset: E. coli AcrB pore subdomain trimer with conserved structural elements (left) and binding pockets (right) for substrate recognition. The flexible loop (aqua) is found at the bottom of PC1/PC2 and entry into the PBP (pink). The PBP is separated from the DBP (gray) by the switch loop (green). The DBP is important for substrate recognition and contains unique properties, typically targeted for EPI development, such as the hydrophobic trap (HT; dark pink). (C) Summary and implications of the Goldilocks binding affinity hypothesis for substrate selection. Molecules with too weak of an interaction with entrance residues do not enter pump and are considered non-substrates (gold). Substrates (green) must bind tight enough for uptake, but not too tight to impede translocation through the PBP/DBP for eventual translocation to the PAP. Too tight binding may result in a molecule that becomes stuck in the DBP, blocking efflux activity and acting as an inhibitor (red).