Fig 1.
Phylogeny of icr-Mo (AXE82_07515).
A. The unrooted nucleotide-based radial phylogram of icr-Mo. Nucleotide sequences of putative sulfatases homologous to MCR-1 are included in the analysis as a reference for evolutionary distance. Four distinct subclades included i) MCR-1/2 variants (in green); ii) icr-Mo (a chromosomally-encoded homologue of mcr-1/2 variants), which are mostly from (but not limited to) the Moraxella species; iii) Sulfatases homologous to MCR-1 (in pink) and iv) MCR-3/4 variants (in Orange). In particular, a chromosomally-encoded colistin resistance determinant, Neisseria eptA falls under the evolutionarily-distant Sulfatase cluster. Unlike ICR-Mo (AXE82_07515), a functional PE transferase (in red), Z1140 is an experimentally-verified non-functional PE transferase (S2 Fig) and acts as an internal reference in this phylogeny. B. Phylogenetic tree of icr-Mo (AXE82_07515) and its close homologs. Two distinct subclades are clustered, including i) MCR-1/2 variants (in the green background) and ii) Chromosomal MCR-like variants (in blue background, with genes from Moraxella species highlighted in red). The tree has been rooted with Z1140, an experimentally-verified non-functional PE transferase (S2 Fig). An asterisk has been used to indicate MCR-1 sequences with a silent DNA point mutation. The nucleotide sequence-based phylogeny of ICR-Mo homologs was inferred using the maximum likelihood method and a GTR nucleotide substitution model. The percentages of replicate trees in which the associated taxa are clustered in the bootstrap test (1000 replicates) is shown next to the branches. A discrete gamma distribution was used to model evolutionary rate differences among sites with some evolutionarily invariable sites. Accession numbers corresponding to the nucleotide sequence used have been indicated in the figure.
Fig 2.
Chemical mechanism for PEA modification of lipid A by EptA/MCR-1/ICR-Mo.
A. Scheme for cleavage reaction of an alternative substrate NBD-glycerol-3-PEA by ICR-Mo (AXE82_07515) into NBD-glycerol and an adduct of ICR-Mo-bound PEA. PEA refers to phosphoethanolamine. NBD was highlighted in magenta, whereas PEA was indicated in red. LC/MS identification of the mixture of the ICR-Mo-mediated hydrolysis reaction (B) and the NBD-glycerol-3-PEA substrate (C). The inside gels denote the TLC-based visualization of the NBD-glycerol-3-PEA substrate (in Panel B) and the ICR-Mo-mediated hydrolytic product, NBD-glycerol (in Panel C). NBD-glycerol-3-PEA appears at m/z of 814.1, whereas the resultant product NBD-glycerol occurs at m/z of 691.5. C. Thin layer chromatography (TLC) detection for the conversion of NBD-glycerol-PEA lipid substrate by the ICR-Mo (AXE82_07515) into NBD-glycerol. D. Transfer of PEA from ICR-Mo-bound PEA to lipid A, generating the PPEA-lipid A product. Position of PPEA depicted is only suggestive. MALDI-TOF-MS evidence for the structural alteration of the lipid A moieties of lipopolysaccharide (LPS) in E. coli expressing EptA (E), MCR-1 (F) and ICR-Mo (AXE82_07515) (G). The peak of the bis-phosphorylated hexa-acylated lipid A varies at m/z of 1796.063 ~ 1797.426, whereas resultant derivative with PEA modification (PPEA-1(or 4’)-lipid A) exhibits at m/z varying from 1919.409 to 1920.087.
Fig 3.
Paralleled PE-recognizing cavities amongst ICR-Mo, EptA and MCR-1.
A. Chemical structure of the PE lipid substrate molecule. B. EptA has a cavity for the entry of the PE lipid substrate. C. A PE-recognizable cavity is present in MCR-1 enzyme. D. A conservative PE-binding cavity is also shared by ICR-Mo (AXE82_07515). A Zn2+-bound five-residues forming motif is conserved in PE lipid substrate-interactive cavities of three enzymes EptA (E), MCR-1 (F) and ICR-Mo (G). Comparative analyses of the seven conserved PE-recognizable residues from EptA (H), MCR-1 (I) and ICR-Mo (J). The enlarged surface structures of PE-bound cavities are consistently generated through molecular docking together with structural modelling. PE molecules are illustrated with red sticks, and cavity is highlighted with an arrow. The photographs are generated using PyMol. The conserved residues are labelled, and also listed in S2 Fig. Designations: PE: phosphatidylethanolamine.
Fig 4.
Functional analyses for the Zinc, PE-interacting motifs of substrate-bound cavities amongst EptA, MCR-1 and ICR-Mo (AXE82_07515).
A. Western blot analyses of expression of ICR-Mo and its 12 point-mutants in E. coli. B. Structure-guided site-directed mutagenesis analyses for the Zn2+-binding motif of ICR-Mo using the LBA plate-based assays of colistin resistance. The five residues (E248, T287, H397, D472 and H473) are required for the binding of ICR-Mo to the zinc ion. C. Functional mapping of the PE-interactive residues of ICR-Mo in the context of colistin resistance. D. The measurement of colistin MIC of the E. coli strains carrying the wild-type icr-Mo (and/or its point-mutants). Level of colistin resistance was tested the LBA plates. A representative result is given from no less than three independent trials. In terms of level of colistin resistance, AXE82_07515 is almost identical to EptA, but only half of that associated with MCR-1. Designations: Vec, an empty vector pBAD24; WT, the wild-type of ICR-Mo (AXE82_07515).
Fig 5.
MALDI-TOF MS identification of LPS-lipid A structure suggests different roles of putative cavity-forming sites in catalysis mechanism of ICR-Mo (AXE82_07515).
MALDI-TOF MS profile of the LPS-lipid A pool isolated from the two negative controls, E. coli strain MG1655 alone (A) and with the empty vector pBAD24 (B). C. Appearance of a unique peak of the mono-modified lipid A, PPEA-1(or 4’)-lipid A, in the positive control, MG1655 strain carrying the wild-type of ICR-Mo (AXE82_07515). The two mutations of N110A (D) and T114A (E) failed to fully inactivate the enzymatic activity of ICR-Mo in that the modified peak, PPEA-lipid A is present. MALDI-TOF-MS analyses suggests that the three mutations of E118A (F), E246A (G) and T287A (H) impairs the function of ICR-Mo. The point-mutant (S332A) of ICR-Mo still possesses partial activity of catalyzing the transfer of PPEA to the 1(4’)-phosphate group of lipid A moieties. The six point-mutants of ICR-Mo are consistently inactive in the enzymatic activity of PEA transferase in vivo, including K335A (J), H397A (K), H402A (L), D472A (M), H473 (N) and H485A (O), respectively. The MS peak of lipid A species in E. coli is shown at m/z of 1796.744~1797.843, whereas its modified form appears at m/z of 1919.865~1920.457, when functional (even partial active) versions of ICR-Mo (AXE82_07515) are present in E. coli.
Fig 6.
Domain-swapping analyses of ICR-Mo.
A. Schematic illustration for domain-swapping designing amongst the three transmembrane enzymes EptA, MCR-1 and ICR-Mo (AXE82_07515). B. Use of Western blotting to detect the expression of icr-Mo and its mosaic versions. C. Functional evaluation of ICR-Mo derivatives in ability of conferring appreciable growth of E. coli on LBA plates with varied levels of colistin. D. Measurement of colistin MIC of E. coli strains expressing ICR-Mo and its hybrid derivatives.
Fig 7.
Use of MALDI-TOF mass spectrometry to evaluate physiological role of chimeric versions of ICR-Mo/EptA/MCR-1 in structural modifications of LPS-lipid A species in E. coli.
MALDI-TOF MS suggests a single peak of intact LPS-lipid A species in the E. coli alone (A) or carrying the empty vector pBAD24 (B). An additional unique MS peak of the PEA-added (lipid A-4’-PEA) appears in the E. coli strains expressing MCR-1 (C), EptA (D) and ICR-Mo (E). The three hybrid enzymes [TM(AXE82)-MCR-1 (F), TM(AXE82)-EptA (G) and TM(EptA)-AXE82 (I)] have no detectable activities in transfer of PEA moiety to the 1(4’)-phosphate position of lipid A GlcN moieties. H. The mosaic enzyme of TM(MCR-1)-AXE82 exhibits partially enzymatic activity in the formation of lipid A-4’-PPEA from lipid A. The MS spectra (in Panels A and B) functioned as negative controls, whereas the positive controls appeared (in Panels C, D and E). The MS peak of lipid A species in E. coli is shown at m/z of 1796.063~1797.653, whereas its modified form, PPEA-1(or 4’)-lipid A is detected at m/z 1919.409~1920.087.
Fig 8.
Colistin-stimulated ROS production is significantly impaired by the presence of colistin resistance-conferring proteins (EptA/MCR-1/ICR-Mo).
A & B Accumulation of hydrogen peroxide is boosted by colistin stress in E. coli. C & D The presence of plasmid-borne EptA inhibits colistin-induced production of hydrogen peroxide in E. coli. E & F Colistin-triggered production of hydrogen peroxide is decreased upon the expression of MCR-1 in E. coli. G & H The expression of icr-Mo greatly impairs colistin-promoted accumulation of hydrogen peroxide in E. coli. The intra-cellular ROS level was detected with an oxidant-sensitive dye, DCFH2-DA. The fluorescent product of DCF was generated due to the oxidation of the dye by hydrogen peroxide. The fluorescence intensity was quantified with a Zeiss LSM 510 Meta confocal laser scanning microscope (100x oil immersion objective). The hydrogen peroxide produced is denoted in green.
Fig 9.
Relative ratio of colistin-induced ROS levels in E. coli carrying icr-Mo (mcr-1 and/or eptA).
Ratio of fluorescent cells was obtained by counting the number of cells with/without fluorescence (illustrated in Fig 8). In every group, over 500 cells counted from 4 individual photographs. The data was assessed using one-way analysis of variance (ANOVA) followed by Tukey–Kramer multiple comparisons post hoc test.
Fig 10.
A working model that ICR-Mo stops colistin-induced hydroxyl radical killing in E. coli.
A. Scheme for ROS production triggered by colistin in E. coli. B. Impairment of colistin-induced ROS formation in icr-Mo-bearing E. coli. C. Chemical rescue experiments reveal that a Fenton reaction is involved in the colistin-activated hydroxyl radical killing pathway in E. coli. The LPS-lipid A moiety refers to an initial target for colistin treatment. Bipyridine is a well-studied ferric chelator, and L-cysteine is the ROS scavenger.