The PmrA-regulated pmrC gene mediates phosphoethanolamine modification of lipid A and polymyxin resistance in Salmonella enterica

Abstract

The PmrA/PmrB regulatory system of Salmonella enterica controls the modification of lipid A with aminoarabinose and phosphoethanolamine. The aminoarabinose modification is required for resistance to the antibiotic polymyxin B, as mutations of the PmrA-activated pbg operon or ugd gene result in strains that lack aminoarabinose in their lipid A molecules and are more susceptible to polymyxin B. Additional PmrA-regulated genes appear to participate in polymyxin B resistance, as pbgP and ugd mutants are not as sensitive to polymyxin B as a pmrA mutant. Moreover, the role that the phosphoethanolamine modification of lipid A plays in the resistance to polymyxin B has remained unknown. Here we address both of these questions by establishing that the PmrA-activated pmrC gene encodes an inner membrane protein that is required for the incorporation of phosphoethanolamine into lipid A and for polymyxin B resistance. The PmrC protein consists of an N-terminal region with five transmembrane domains followed by a large periplasmic region harboring the putative enzymatic domain. A pbgP pmrC double mutant resembled a pmrA mutant both in its lipid A profile and in its susceptibility to polymyxin B, indicating that the PmrA-dependent modification of lipid A with aminoarabinose and phosphoethanolamine is responsible for PmrA-regulated polymyxin B resistance.

FIG. 1.

(A) Schematic representation of the pmrCAB operon in wild-type Salmonella and in mutants with a partial (ΔpmrC1 and ΔpmrC1.1) or complete (ΔpmrC2) deletion of the pmrC open reading frame. (B) β-Galactosidase activity (in Miller units) expressed by strains harboring a chromosomal lac transcriptional fusion to the PmrA-activated pbgP gene that were grown logarithmically in N-minimal medium, pH 5.8, with 10 μM MgCl₂. Transcription was investigated in wild-type (14028s), ΔpmrC1 (EG13927), and ΔpmrC2 (EG13633) genetic backgrounds. Data correspond to mean values from three independent sets of experiments performed in duplicate. Transcription of the PmrA-activated pbgP gene was similar in the wild-type and ΔpmrC1 strains, but it was decreased in the ΔpmrC2 mutant. (C) Western blot analysis of cell extracts prepared from the ΔpmrC1.1 mutant (EG14592) containing the ppmrCFLAG plasmid, which expresses the pmrCflag gene from its own promoter, after logarithmic growth in N-minimal medium, pH 7.7, with 10 μM (L) or 10 mM (H) MgCl₂. The total protein from equal amounts of bacterial cells, as adjusted by the OD₆₀₀, was run in an SDS-10% polyacrylamide gel, transferred onto a nitrocellulose membrane, and developed by using anti-FLAG antibodies. The ΔpmrC1.1 mutant displays normal PmrA regulation, as the PmrC-FLAG protein is produced by bacteria grown in a low Mg²⁺ concentration but is not detected when bacteria are grown in a high Mg²⁺ concentration.

FIG. 2.

Lipid A species profiles from wild-type (14028s) (A), pbgP (EG9241) (B), ΔpmrC1.1 (EG14590) (C), ΔpmrC1.1/ppmrC (EG14595) (D), and ΔpmrC1.1/vector (EG14656) (E) strains grown to logarithmic phase in N-minimal medium, pH 5.8, with 10 μM MgCl₂, and analyzed by negative-ion-mode MALDI-TOF mass spectrometry. These profiles show that the pmrC mutant lacks lipid A species modified with phosphoethanolamine.

FIG. 3.

(A) Polymyxin B killing assay of wild-type (14028s), ΔpmrC1.1 (EG14590), ΔpmrC1.1/vector (EG14656), ΔpmrC1.1/ppmrC (EG14595), and pmrA (EG7139) strains grown to logarithmic phase in N-minimal medium, pH 5.8, with 10 μM MgCl₂. Polymyxin B was added to a final concentration of 10 μg/ml, and the bacteria were incubated for 1 h at 37°C. The samples were diluted in PBS and plated on LB agar plates to determine the numbers of CFU. Survival values given are relative to those of PBS-treated samples. The ΔpmrC1.1 (EG14590) and ΔpmrC1.1/vector (EG14656) strains were significantly more sensitive to polymyxin B than was the wild-type (14028s) strain (P < 0.01). The complemented strain ΔpmrC1.1/ppmrC (EG14595) was significantly more resistant to polymyxin B than were strains ΔpmrC1.1 (EG14590) and ΔpmrC1.1/vector (EG14656) (P < 0.01). (B) Polymyxin B killing assay of wild-type (14028s), ΔpmrC1 (EG13927), pbgP (EG9241), pbgP ΔpmrC1 (EG14372), and pmrA (EG7139) strains grown and tested as described above, except that polymyxin B was added at final concentrations of 1 and 5 μg/ml. The difference in the polymyxin B (1 μg/ml) susceptibilities of strains pbgP ΔpmrC1 (EG14372) and pmrA (EG7139) was not statistically significant (P = 0.7), indicating that the pbgP and pmrC loci mediate PmrA-controlled polymyxin B resistance. (C) Polymyxin B killing assay of wild-type (14028s), pmrA505 (EG9492), pmrA505 ΔpmrC1.1 (EG14368), pmrA505 pbgP (EG9868), pmrA505 pbgP ΔpmrC1.1 (EG14369), and pmrA (EG7139) strains grown and tested as described for panel A, except that polymyxin B was added at 1, 5, and 20 μg/ml. Note the logarithmic scale (a linear scale is used in the insets) on the y axis. The data correspond to mean values from three independent sets of experiments performed in duplicate. The data demonstrate that the inactivation of the pmrC gene increases the susceptibility of cells to polymyxin B and that a pbgP ΔpmrC1 double mutant exhibits the same level of polymyxin B susceptibility as the pmrA null mutant.

FIG. 4.

Lipid A species profiles for the pmrA505 (EG9492) (A), pmrA505 pbgP ΔpmrC1.1 (EG14369) (B), pbgP ΔpmrC1 (EG14372) (C), and pmrA (EG7139) (D) strains grown to logarithmic phase in N-minimal medium, pH 5.8, with 10 μM MgCl₂, and analyzed by negative-ion-mode MALDI-TOF mass spectrometry. These profiles show that the pbgP ΔpmrC1 and pmrA505 pbgP ΔpmrC1.1 mutants have the same lipid A profile as the pmrA null mutant.

FIG. 5.

(A) Western blot analysis of inner and outer membranes prepared from the ΔpmrC1.1 strain containing either the pBAC108L vector (EG14656) or the ppmrCFLAG plasmid (EG14592), which carries a pmrC gene directed by its own promoter and expresses a PmrC protein tagged with a FLAG epitope at its C terminus. Bacteria were grown to the logarithmic phase in N-minimal medium, pH 7.7, with 10 μM MgCl₂. Inner and outer membranes were prepared by sucrose density gradient centrifugation. Twenty micrograms of protein from the inner and outer membranes was boiled for 10 min, run in an SDS-10% polyacrylamide gel, transferred onto a nitrocellulose membrane, and developed by using anti-FLAG antibodies. To examine the purity of the membrane preparations, we determined the NADH oxidase activity by measuring the oxidation of NADH at 340 nm, and these values are expressed as follows: 100 × μmol of substrate oxidized/min/mg of protein. The analysis demonstrates that the PmrC protein localizes to the inner membrane. (B) Kyte-Doolittle hydropathy plot (25) of the PmrC protein generated by DNA Strider 1.3 software. (C) The left panel shows the predicted topology of the PmrC protein. The numbers correspond to the positions in the PmrC protein at which in-frame fusions were generated to the PhoA and LacZ proteins. The right panel shows alkaline phosphatase and β-galactosidase activities displayed by the phoN strain (EG14286) harboring plasmids pPmrC₁₅₀-lacZ′, pPmrC₁₅₀-phoA′, pPmrC₁₈₁-lacZ′, pPmrC₁₈₁-phoA′, pPmrC₂₉₅-lacZ′, and pPmrC₂₉₅-phoA′ when streaked onto LB agar plates containing either XP (40 μg/ml) or X-Gal (40 μg/ml). These data suggest that the C-terminal region (amino acids 177 to 547) of PmrC localizes to the periplasm.