Mobile Colistin Resistance Enzyme MCR-3 Facilitates Bacterial Evasion of Host Phagocytosis

Abstract

Mobile colistin resistance enzyme MCR-3 is a phosphoethanolamine transferase modifying lipid A in Gram-negative bacteria. MCR-3 generally mediates low-level (≤8 mg L^-1 ) colistin resistance among Enterobacteriaceae, but occasionally confers high-level (>128 mg L^-1 ) resistance in aeromonads. Herein, it is determined that MCR-3, together with another lipid A modification mediated by the arnBCADTEF operon, may be responsible for high-level colistin resistance in aeromonads. Lipid A is the critical site of pathogens for Toll-like receptor 4 recognizing. However, it is unknown whether or how MCR-3-mediated lipid A modification affects the host immune response. Compared with the wild-type strains, increased mortality is observed in mice intraperitoneally-infected with mcr-3-positive Aeromonas salmonicida and Escherichia coli strains, along with sepsis symptoms. Further, mcr-3-positive strains show decreased clearance rates than wild-type strains, leading to bacterial accumulation in organs. The increased mortality is tightly associated with the increased tissue hypoxia, injury, and post-inflammation. MCR-3 expression also impairs phagocytosis efficiency both in vivo and in vitro, contributing to the increased persistence of mcr-3-positive bacteria in tissues compared with parental strains. This study, for the first time, reveals a dual function of MCR-3 in bacterial resistance and pathogenicity, which calls for caution in treating the infections caused by mcr-positive pathogens.

Keywords: colistin; lipid A modification; mcr-3; phagocytosis; virulence.

Figure 1

Phenotype and colistin resistance mechanism of the mcr‐3‐positive AS1 strain. A) MICs of colistin against A. salmonicida AS1 and E. coli DH5α harboring the recombinant mcr‐3‐carrying plasmid or empty vector. B) Comparison of the survival rates (%) of strain AS1 with/without mcr‐3 in the presence of different concentrations of LL‐37 (two‐tailed unpaired t‐test; ^* P < 0.05). C) Growth curve of strain AS1 with/without mcr‐3. The mcr‐3‐positive strain was cultivated in the presence of colistin at concentrations of 0, 2, 4, or 8 mg L⁻¹. D) TEM micrographs of AS1 with/without mcr‐3. The mcr‐3‐positive strains were treated with colistin at concentrations of 1, 8, 32 (1/2‐fold of MIC), 64 (1‐fold of MIC), or 128 mg L⁻¹ (2‐fold of MIC). E) Negative‐ion MALDI‐TOF mass spectrometry analysis of strain AS1 carrying plasmid pHSG299. F) Negative‐ion MALDI‐TOF mass spectrometry analysis of strain AS1 carrying plasmid mcr‐3‐pHSG299. G) Chemical structure analysis of modified lipid A from strain AS1. H) Mass spectrometry analysis of Aeromonas caviae harboring plasmid pT. The structure shows the arnT‐mediated lipid A modification.

Figure 2

mcr‐3‐positive bacteria showed increased pathogenicity in mice compared with the mcr‐3‐negative strain. A) Pathological sections of lung removed from dead mice. B) Clinical scores of infected mice. The overall status of each mouse was assessed at each time point. C) Body temperatures of infected mice (three mice were tested at each time point; ^*** P < 0.001). D) Bacterial counts per 100 µL of blood (three mice were tested at 3 h postinfection; ^* P < 0.05). E) Enumeration of monocytes and neutrophils per liter of blood (three mice were tested at 6 h postinfection; ^* P < 0.05). F) Levels of IL‐6, IL‐1β, and TNF‐α expression in serum at 3 and 12 h postinfection (three mice were tested at each time point; ^* P < 0.05). G) Enumeration of monocytes and neutrophils per liter of blood (three mice were tested at 6 h postinfection; ^* P < 0.05). H) Clinical scores of mice. The overall status of each mouse was assessed at each time point. I) Body temperatures of mice (three mice were tested at each time point; ^** P < 0.01 and ^*** P < 0.001). Statistical significance was assessed by two‐tailed unpaired t‐test.

Figure 3

Bacterial loads in the tissues of infected mice. A,C) Bacterial counts per 100 µL of blood and peritoneal washes (three mice were tested at each time point). B,D) Bacterial counts in 100 µL of tissue homogenate (three mice were tested at each time point). Results indicated the mean ± SEM. Statistical significance was assessed by two‐tailed unpaired t‐test; ^** P < 0.01 and ^* P < 0.05.

Figure 4

A) Cross‐sections of lung, kidney, and heart tissues collected 24 h postinfection with mcr‐3‐positive or mcr‐3‐negative E. coli. Sections were immunostained with anti‐HIF‐1α (green fluorescence indicates areas of hypoxia, bar = 10 µm). B) Quantification of hypoxic areas following immunostaining of tissue sections (calculated from at least five fields per slide, two‐tailed unpaired t‐test; ^** P < 0.01).

Figure 5

A) Cross‐sections of liver, spleen, and kidney tissues 24 h postinfection with mcr‐3‐positive or mcr‐3‐negative E. coli. Sections were immunostained with anti‐MAC387 (red fluorescence, monocytes/macrophages) and anti‐E. coli (green fluorescence, E. coli, bar = 10 µm). B) Percentages of red and green double positive area in red area were quantificated in tissue sections, indicative of macrophages containing phagocytosed E. coli (calculated from at least five fields per slide; ^*** P < 0.001). C) Quantitative analysis of absolute counts of liver inflammatory cells by flow cytometry (n = 3) following infection with mcr‐3‐positive or mcr‐3‐negative E. coli harboring a GFP fluorescent label. D) Absolute counts of neutrophils, lymphocytes, and macrophages in liver containing phagocytosed fluorescently tagged E. coli at 24 h postinfection (n = 3, ^** P < 0.01). E) Quantitative analysis of percentages and absolute counts of inflammatory cells harboring GFP‐tagged mcr‐3‐positive or mcr‐3‐negative E. coli in the spleen, determined by flow cytometry (n = 3). F) Absolute counts of neutrophils, lymphocytes, and macrophages in spleen tissues containing phagocytosed fluorescently tagged E. coli at 24 h postinfection (n = 3, ^* P < 0.05, ^** P < 0.01). Statistical significance was assessed by two‐tailed unpaired t‐test.

Figure 6

Compared with the wild‐type strain, mcr‐3‐positive AS1 suffered significantly less phagocytosis and showed reduced macrophage stimulation. A) Representative confocal laser scanning microscopy images showing phagocytosis of AS1 strains containing plasmids pHSG299 or pHSG299‐mcr‐3, along with uninfected controls. Calculation of the phagocytosis index (engulfed bacteria per macrophage, calculated from at least five fields per slide) based on confocal microscopy images (≥200 cells were scored per well). B) Representative confocal laser scanning microscopy images showing phagocytosis of DH5α strains containing plasmids pHSG299 or pHSG299‐mcr‐3, along with uninfected controls. Calculation of the phagocytosis index (engulfed bacteria per macrophage, calculated from at least five fields per slide) based on confocal microscopy images (≥200 cells were scored per well). C) Transcription of IL‐1β and TNF‐α in RAW264.7 macrophages stimulated with mcr‐3‐positive or mcr‐3‐negative AS1. D) Transcription of IL‐1β and TNF‐α in RAW264.7 macrophages stimulated with MCR‐3‐modified/unmodified LPS extracted from AS1. E) Immunoblotting analysis. Total expression of NF‐κB‐pathway proteins (phosphorylated‐p65, p65, β‐actin, and IκBα) in the lysates of RAW264.7 macrophages treated for 60 min with modified/unmodified LPS purified from mcr‐3‐positive/negative AS1. F) Nuclear protein expression in RAW264.7 macrophages treated for 60 min with MCR‐3‐modified/unmodified LPS from AS1. Histone was used as a control. Statistical significance was assessed by two‐tailed unpaired t‐test; ^** P < 0.01 and ^* P < 0.05. Error bars represent means ± SEM from triplicate wells.