Article Figures & Data
Figures
- FIG 1
Schemes of chemical mechanism and colistin resistance of MCR-1, MCR-2, and EptA. (A) Chemical reaction for the transfer of PEA to lipid A by MCR-2. MCR-2 catalyzes the addition of PEA to position 1 or 4′ of lipid A, giving the final products PEA-1(4′)-lipid A and diacylglycerol. The enzyme (MCR-1 or -2 or EptA) is depicted with a ribbon structure comprising the N-terminal catalytic domain (in green) and the transmembrane domain (in blue) at the C terminus. The chemical structures of molecules are illustrated with ChemDraw software. (B) The first half reaction of MCR-1 or -2 or EptA is defined by the removal of PEA from PE, giving the final product DG and the intermediate product enzyme-bound PEA. (C) The second half reaction of MCR-1 or -2 or EptA comprises the generation of the final product Kdo2 [di(3-deoxy-d-manno-octulosonic acid)]-lipid A-4′-PPEA through the transfer of PEA from the adduct of enzyme-bound PEA to the recipient Kdo2-lipid A. (D) Cartoon of the model proposed for the bacterial surface structure of colistin-sensitive E. coli. (E) Working model for structural modification of lipid A anchored on bacterial surface involved in colistin resistance. PE, phosphatidylethanolamine; DG, diacylglycerol; PEA, phosphoethanolamine; OM, outer membrane; IM, inner membrane. (F) Chemical structure of the cationic antibiotic peptide, colistin. Positively charged elements are denoted in red.
- FIG 2
Identification of a conserved PE substrate-binding cavity among EptA and MCR-1 and -2. (A) Chemical structure of the detergent DDM, an analogue of the PE lipid substrate for MCR-1 and -2. (B) Chemical structure of the PE molecule, the lipid substrate of MCR-1 and -2. (C to H) Surface structure illustrations of the DDM-bound cavities in EptA (C), MCR-1 (E), and MCR-2 (G) and comparison of surface structures of the PE-bound cavities in EptA (D), MCR-1 (F), and MCR-2 (H). DDM/PE molecules are illustrated with blue sticks, and cavity is highlighted with an arrow. Images were created using PyMol. DDM, n-dodecyl-β-d-maltoside; PE, phosphatidyl ethanolamine.
- FIG 3
Phylogeny of MCR-like enzymes. All the MCR-like proteins used here were sampled from the protein database of the NCBI website and subjected to phylogenetic analyses using Clustal Omega (https://www.ebi.ac.uk/Tools/msa/clustalo/) (67). The final phylogenetic output was visualized with TreeView (68). The phylogenetic analysis indicates that the MCR-like enzymes are categorized into two groups: subclade I, the MCR-1 and -2-like enzymes, and subclade II, the non-MCR-1 and -2-like enzymes, such as EptA. Subclade I includes 9 MCR-1 variants, 3 MCR-2 variants, and putative MCR-1/-2 progenitors from five different Moraxella species (51, 69). The product of the Z1140 locus (in gray) of E. coli O157:H7, a member of the PEA lipid A transferases lacking a role in colistin resistance, is used as an internal reference.
- FIG 4
Structure-guided functional analyses of the PE-binding cavity of MCR-2. (A) Surface structure of MCR-2 with the cavity required for entry and binding of PE substrate. (B) Enlarged view of the 5-residue Zn2+-binding motif. (C) Surface structure of MCR-2 in the counter clockwise rotation of 30 degrees with fine-structural illustration of the cavity for entry of the PE substrate. (D) Enlarged view of the 7-residue motif with essential roles in the PE-binding cavity of MCR-2. The 7 amino acids (N106, T110, E114, S328, K331, H393, and H476) are proposed to participate in the formation of the PE cavity of MCR-2. (E) Western blotting-based expression analyses of MCR-2 and its 12 point-mutants in E. coli. Given the limit of the wells of PAGE (10 per gel), the photograph of Western blotting here was generated through a combination of two different gel images in which the protein samples were separated. (F) Site-directed mutagenesis assay of the Zn2+-binding motif in the context of MCR-2 colistin resistance. The 5 residues in the Zn2+-binding motif of MCR-2 are E244, T283, H388, D463, and H464. (G) Site-directed mutagenesis assay of the PE-binding cavity in the context of MCR-2 colistin resistance. The 2 periplasmic-facing helices (in light golden in panel D) possess 3 crucial residues (namely, N106, T110, and E114) that play roles in the binding of MCR-2 to the PE substrate molecule. (H) Comparison of colistin MICs in E. coli strains carrying either wild-type mcr-2 or its point mutants. (I) TLC-based assays of the enzymatic activities of MCR-2 and its 12 point mutants. Structure-guided site-directed mutagenesis was performed as recommended by the manufacturer. Vec, empty-vector-bearing strain; WT, wild type; Ori, origin. All strains tested here are listed in Table S1.
- FIG 5
MCR-1 and -2 (and EptA) can release PEA (phosphoethanolamine) from the alternative substrate NBD-glycerol-3-PEA. (A) Scheme for hydrolysis of the alternative substrate NBD-glycerol-PEA into NBD-glycerol and the enzyme-bound PEA adduct. (B) LC-MS identification of NBD-glycerol-3-PEA. The inset gel shows NBD-glycerol-3-PEA separated with thin-layer chromatography (TLC). (C) LC-MS identification of the enzymatic activities of MCR-1/MCR-2/EptA that catalyze cleavage of NBD-glycerol-PEA lipid substrate into NBD-glycerol. The inset gel shows TLC analysis of MCR-1-catalyzed hydrolysis of NBD-glycerol-PEA lipid substrate, giving its product, NBD-glycerol.
- FIG 6
Comparative analyses of hydrolytic activities of MCR-2 and its derivatives. (A) Molecular designs for hybrid versions of MCR-1 and -2 and EptA. +, functional; −, nonfunctional. (B) Colistin MICs of E. coli MG1655 strains expressing mcr-1 and -2 and their derivatives. (C) Western blot (WB) analysis of the expression of mcr-1 and -2 and their derivatives. Of note, QTOF MS and circular dichroism analyses show that the different migration rates of the chimeric proteins (especially TM-MCR-1 and TM1-EptA) in SDS-PAGE are due not to protein degradation or misfolding but to differing charges. (D) TLC assays of the enzymatic activities of MCR-1 and -2 and their hybrid versions in vitro. Domain-swapping analyses suggest that domains are functionally exchangeable between MCR-1 and MCR-2 but not between EptA and MCR-1 or MCR-2. This is consistent with the fact that EptA is evolutionarily distant from MCR-1 and -2, as illustrated in the phylogeny of MCR-like proteins (Fig. 3). TM, transmembrane region of EptA; TM1, transmembrane region of MCR-1; TM2, transmembrane region of MCR-2; PEA, phosphoethanolamine; TLC, thin-layer chromatography; gly, glycerol; NBD-gly-3-PEA, fluorescent label (1-acyl-2-{12-[(7-nitro-2-1,3-benzoxadiazol-4-yl) amino] dodecanoyl}-n-glycerol-3-phosphoethanolamine); Ori, origin.
- FIG 7
MS determination of altered structures of LPS-lipid A anchored on the bacterial surface upon the expression of MCR-1 and -2 and its derivatives. (A) Silver-staining analysis of LPS-lipid A isolated from diverse E. coli strains with or without mcr-1 or -2 or their derivatives. +, lipid A with the addition of PEA; −, intact lipid A; Vec, empty vector. (B, C) LPS-lipid A profiles of the two negative-control strains, E. coli MG1655 alone (B) and MG1655 with the empty vector pBAD24 (C). (D to F) Appearance of a unique peak for the modified lipid A (PPEA-4′-lipid A) in the E. coli strains expressing any one of the three genes eptA (D), mcr-1 (E), and mcr-2 (F). (G, H) No modification of lipid A by the expression of the MCR-1 hybrid (TM-MCR-1) whose transmembrane region is replaced by its counterpart from EptA (G) or the reverse hybrid, TM1-EptA (H). (I, J) MALDI-TOF MS suggests only an unmodified lipid A peak is present in the MG1655 strain regardless of the presence of TM-MCR-2 (I) or TM2-EptA (J). (K, L) Identification of a modified lipid A, PPEA-4′-lipid A, in the E. coli strains harboring either TM1-MCR-2 (K) or TM2-MCR-1 (L). LPS, lipopolysaccharide; TM-MCR-2, a derivative of MCR-2 in which the transmembrane region is replaced by the counterpart from EptA; TM2-EptA, a hybrid version of EptA carrying the transmembrane region from MCR-2; TM1-MCR-2, a derivative of MCR-2 with the MCR-1 transmembrane region; TM2-MCR-1, a derivative of MCR-1 with the MCR-2 transmembrane region. The MS peak of lipid A species in E. coli occurs at m/z of 1,796.170 to 1,797.256, whereas it appears at m/z 1,919.615 to 1,920.315 in E. coli cells with functional expression of MCR-1 and -2, EptA, and its derivatives, because the modified PEA species [PPEA-1(4′)-lipid A] is present. Here, the PEA mass is 123 unified atomic mass units (u).
Tables
- TABLE 1
MALDI-TOF MS profiles of lipid A species from E. coli MG1655 strains expressing eptA, mcr-1, mcr-2, and their point mutants
Protein,
lipid speciesMass (m/z) of lipid species from E. coli MG1655 bearing: Empty
vectorNo vector
(WTa)Vector expressing protein with indicated mutation EptA N106A T110A E114A E240A T280A S325A K328A H378A H383A D452A H453A H465A Lipid A 1,796.700 1,797.936 1,797.02 1,797.083 1,796.950 1,797.069 1,796.719 1,797.138 1,797.005 1,796.893 1,796.734 1,796.935 1,796.814 1,796.869 PPEA-4′-lipid A 1,919.983 1,920.077 1,920.141 1,920.195 MCR-1b N108A T112A E116A E246A T285A S330A K333A H390A H395A D465A H466A H478A Lipid A 1,796.567 1,796.915 1,796.572 1,796.707 1,796.598 1,796.645 1,796.521 1,796.651 1,796.676 1,796.786 1,796.535 1,796.675 1,796.718 1,796.912 PPEA-4′-lipid A 1,919.969 1,919.614 1,919.753 1,919.695 MCR-2 N106A T110A E114A E244A T283A S328A K331A H388A H393A D463A H464A H476A Lipid A 1,796.243 1,796.939 1,796.986 1,796.812 1,796.622 1,796.571 1,796.722 1,796.717 1,796.538 1,796.860 1,796.070 1,796.611 1,796.602 1,796.664 PPEA-4′-lipid A 1,919.991 1,919.114 1,919.857 1,919.753
FIG S1
Sequence alignment of three colistin resistance genes MCR-1, MCR-2, and NgEptA. The transmembrane region, underlined in blue, is revealed through prediction with TMHMM server version 2.0 (http://www.cbs.dtu.dk/services/TMHMM/). The multiple sequence alignment was carried out using Clustal Omega (https://www.ebi.ac.uk/Tools/msa/clustalo/), with the current output by the program ESPript 3.0 (http://espript.ibcp.fr/ESPript/cgi-bin/ESPript.cgi) (61). The five Zn2+-interacting residues are highlighted with red arrows, and the seven PE substrate-binding amino acids are indicated with blue arrows. Identical residues are in white letters with red background, similar residues are in red letters with white background, and varied residues are in black letters. The protein secondary structure is illustrated in cartoon form (on top). α, α-helix; β, beta-sheet; T, turn; η, coil; TMH, transmembrane helix; PH, periplasmic-facing helix; Ng, Neisseria gonorrhoeae. Download FIG S1, JPG file, 0.6 MB.
Copyright © 2018 Xu et al.
This content is distributed under the terms of the Creative Commons Attribution 4.0 International license.
FIG S2
Structural reanalysis of Neisseria gonorrhea EptA defines a functional cavity for the entry and binding of PE substrate molecules. (A) Schematic illustration of surface structure of N. gonorrhea EptA (NgEptA). (B) Enlarged view of a 5-residue motif implicated in Zn2+ binding. (C) Surface structure of NgEptA rotated 35°C in the reverse direction. (D) Enlarged view of a 7-residue motif having a role in the PE substrate-binding cavity of NgEptA. The 7 amino acids involved into the PE substrate-binding cavity of NgEptA are N106, T110, E114, S325, K328, H383, and H465. (E) Western blot analyses of the point mutants of EptA with anti-6×His rabbit antiserum as primary antibody. (F) Functional mapping of the Zn2+-binding motif using assays of colistin resistance on LBA plates. The five residues (E240, T280, H378, D452, and H453) are required for Zn2+ binding of NgEptA. (G) Site-directed mutagenesis-based analyses of the PE-binding cavity in the context of colistin resistance conferred by NgEptA. The two periplasmic-facing helices (in light pink) participate in the binding of NgEptA to PE substrate molecules in that they carry the three essential amino acids (namely, N106, T110, and E114). (H) Comparative analyses of colistin MICs in the E. coli strains carrying either the wild-type NgEptA or its point mutants. Levels of colistin resistance were assayed using LBA plates (F and G). A representative result from no less than three independent trials is given. Vec, empty pBAD24 vector. Colistin MIC trials were conducted using the micro-broth dilution method, and the breakpoint was set according to the guidelines of the European Committee on Antimicrobial Susceptibility Testing (EUCAST 2015, version 5.0) (70). Ng, Neisseria gonorrhoeae. Download FIG S2, JPG file, 0.5 MB.
Copyright © 2018 Xu et al.
This content is distributed under the terms of the Creative Commons Attribution 4.0 International license.
FIG S3
Structural dissection of MCR-2 defines a conserved cavity required for the PE substrate binding. (A) Topological illustration of the intramembrane protein. (B) Overall structure of the integral membrane protein MCR-2 at full length. (C) Surface structure of the complex of MCR-2 and its PE substrate. (D) Enlarged view of the PE-binding cavity in MCR-2. The catalytic domain is in green, the TM region is in blue, the two helices (PH2 and PH2′) are in light pink, and PE substrate is indicated with sticks. The majority of the cavity region is formed by the PH2 and PH2′ helices plus the TM domain (highlighted with a red arrow). The rectangle with pink background refers to the inner membrane layer. Overall structure of MCR-2 was modeled with the structural template of Neisseria meningitidis EptA (PDB code 5FGN) and depicted in ribbon form using PyMol. TM, transmembrane; PH, periplasmic-facing helices; H, α-helices; S, β-sheet; N, N terminus; C, C terminus; PE, phosphatidylethanolamine. Download FIG S3, JPG file, 0.5 MB.
Copyright © 2018 Xu et al.
This content is distributed under the terms of the Creative Commons Attribution 4.0 International license.
FIG S4
Circular dichroism-based secondary structures of EptA, MCR-1 and -2, and their derivatives. (A) CD spectrum of EptA protein. (B) CD profile of MCR-1. (C) CD analyses of MCR-2. (D) Use of CD to visualize secondary structure of EptA derivative, TM1-EptA. (E) CD spectrum of MCR-1 derivative TM-MCR-1. (F) CD-based elucidation of EptA derivative TM2-EptA. (G) CD-aided identification of MCR-2 derivative TM-MCR-2. CD, circular dichroism; TM1-EptA, a derivative of EptA whose TM region is replaced with its counterpart in MCR-1 (panel B); TM-MCR-1, a derivative of MCR-1 whose TM region is replaced with its counterpart in EptA (panel A); TM2-EptA, a derivative of EptA whose TM region is replaced with its counterpart in MCR-2 (panel C); TM-MCR-2, a derivative of MCR-2 whose TM region is replaced with its counterpart in EptA (panel A). Download FIG S4, JPG file, 0.9 MB.
Copyright © 2018 Xu et al.
This content is distributed under the terms of the Creative Commons Attribution 4.0 International license.
FIG S5
ICP-MS analyses of Zn2+ in EptA, MCR-1 and -2, and the four chimeric derivatives (TM-MCR-1, TM1-EptA, TM2-EptA, and TM-MCR-2). Binding of zinc to diversified MCR-like proteins was measured using inductively coupled plasma mass spectrometry (ICP-MS) as described by Loeschner et al. (57). ZnCl2 was used as the positive control. Download FIG S5, TIF file, 2 MB.
Copyright © 2018 Xu et al.
This content is distributed under the terms of the Creative Commons Attribution 4.0 International license.
TABLE S1
Strains and plasmids in this study. Download TABLE S1, DOC file, 0.1 MB.
Copyright © 2018 Xu et al.
This content is distributed under the terms of the Creative Commons Attribution 4.0 International license.
TABLE S2
Primers used in this study. Download TABLE S2, DOC file, 0.1 MB.
Copyright © 2018 Xu et al.
This content is distributed under the terms of the Creative Commons Attribution 4.0 International license.
