Deciphering the molecular and functional basis of TMexCD1: the plasmid-encoded efflux pump of resistance-nodulation-division superfamily

Abstract

Horizontal gene transfer has been demonstrated to be an important driver for the emergency of multidrug-resistant pathogens. Recently, a transferable gene cluster tmexCD1-toprJ1 of the resistance-nodulation-division (RND) superfamily was identified in the plasmids of animal-derived Klebsiella pneumoniae strains, with a higher efflux capacity for various drugs than the Escherichia coli AcrAB-TolC homolog system. In this study, we focused on the differences in the inner membrane pump of these two systems and identified some key residues that contribute to the robust efflux activity of the TMexCD1 system. With the aid of homologous modeling and molecular docking, eight residues from the proximal binding pocket (PBP) and nine from the distal binding pocket (DBP) were selected and subjected to site-directed mutagenesis. Several of them, such as S134, I139, D181, and A290, were shown to be important for substrate binding in the DBP region, and all residues in PBP and DBP showed certain substrate preferences. Apart from the conservative switch loop (L613-623^TMexD1) previously identified in the E. coli AcrB (EcAcrB), a relatively unconservative loop (L665-675^TMexD1) at the bottom of PBP was proposed as a critical element for the robust activity of TMexD1, due to variations at sites E669, G670, N673, and S674 compared to EcAcrAB, and the significantly altered efflux activity due to their mutations. The conservation and flexibility of these key factors can contribute to the evolution of the RND efflux pumps and thus serve as potential targets for developing inhibitors to block the widespread of the TMexCD1 system.

Keywords: RND efflux pump; TMexD1; loop 665–675; multidrug resistance; substrate-binding sites.

Fig 1

Multiple sequence alignments of amino acid sequences of TMexD1 and its homologs. TMexD1 and MexD sequences are from P. aeruginosa, AcrB sequence is from the E. coli AcrB (EcAcrB). The approximate location of the TMexD1 porter domain was marked as arrows under the sequences. The residues selected for mutation were marked as numbers, amino acids within DBP are boxed in pink, and residues within PBP are boxed in red.

Fig 2

Molecular docking of TMexD1 with minocycline. (A) Docking conformations of the distal binding pocket. (B) Docking conformations of the proximal binding pocket. The numbers of each molecular conformation and the corresponding binding energy (kcal/mol) were shown in the upper left corner in an order of energy from low to high, residues involved in hydrogen bonding were labeled, and the interactions were shown as yellow dotted lines.

Fig 3

Drug-resistant for E. coli/TMexD1 and its variants. The MICs for tigecycline (A), tetracycline (B), ciprofloxacin (C), gentamicin (D), minocycline (E), and doxorubicin (F) were measured in the Müller-Hinton (MH) medium. E. coli MG1655ΔacrABΔacrEF strain harboring the empty pEASY-Blunt vector was used as the control. AcrB represents the E. coli MG1655ΔacrABΔacrEF strain carrying the pEASY-Blunt/AcrB plasmid. Mutations within the DBP region are marked in pink boxes and PBP mutations in blue boxes. All experiments were performed in triplicates.

Fig 4

Fluorescence efflux and microscopy of E. coli/TMexD1 and its mutants. (A) Mutations at DBP and (B) mutations at PBP. Control strain only harbors the pEASY-Blunt vector. All experiments were conducted in triplicate and the X-bar was removed for a better view. (C) Fluorescence microscopy using doxorubicin, samples were collected at 0 min and 50 min for each strain and the representative field of view was chosen. Bar = 10 µm.

Fig 5

Structural fitting of the extrusion (cyan) and binding (green) states of TMexD1 homologous model. Subdomains were indicated, and important loops (611–621 and 665–675) were shown with different colors (brown for the binding state and magenta for the extrusion state).