Membrane proteases and aminoglycoside antibiotic resistance

Abstract

We present genetic studies that help define the functional network underlying intrinsic aminoglycoside resistance in Pseudomonas aeruginosa. Our analysis shows that proteolysis, particularly that controlled by the membrane protease FtsH, is a major determinant of resistance. First, we examined the consequences of inactivating genes controlled by AmgRS, a two-component regulator required for intrinsic tobramycin resistance. Three of the gene products account for resistance: a modulator of FtsH protease (YccA), a membrane protease (HtpX), and a membrane protein of unknown function (PA5528). Second, we screened mutations inactivating 66 predicted proteases and related functions. Insertions inactivating two FtsH protease accessory factors (HflK and HflC) and a cytoplasmic protease (HslUV) increased tobramycin sensitivity. Finally, we generated an ftsH deletion mutation. The mutation dramatically increased aminoglycoside sensitivity. Many of the functions whose inactivation increased sensitivity appeared to act independently, since multiple mutations led to additive or synergistic effects. Up to 500-fold increases in tobramycin sensitivity were observed. Most of the mutations also were highly pleiotropic, increasing sensitivity to a membrane protein hybrid, several classes of antibiotics, alkaline pH, NaCl, and other compounds. We propose that the network of proteases provides robust protection from aminoglycosides and other substances through the elimination of membrane-disruptive mistranslation products.

Fig. 1.

Mutations in three AmgRS-regulated genes increase tobramycin sensitivity. Overnight cultures of the indicated strains were diluted 100-fold and spotted on LB-MOPS (pH 7.6) agar containing tobramycin. Images were recorded after 24 h at 37°C. The increased sensitivity of the triple mutant relative to the amgR mutant presumably reflects the basal transcription of htpX, yccA, and/or PA5528 in the mutant background. The following strains were tested (see Table S1 in the supplemental material): AHP36, AHP83, AHP124, AHP125, AHP147, AHP146, AHP145, AHP127, and AHP126.

Fig. 2.

Complementation of tobramycin-sensitive strains. The tobramycin sensitivities of strains carrying plasmid clones of the indicated genes were assessed on LB-MOPS (pH 7.6) agar containing carbenicillin (150 μg/ml). MICs were confirmed in duplicate assays. The IS control carries a transposon insertion in PA3303 which does not affect tobramycin resistance.

Fig. 3.

Toxicity of a MexF-PhoA hybrid protein toward tobramycin-sensitive mutants. The effect of the expression of a toxic (allele L18) or nontoxic (allele I17) MexF-PhoA hybrid on the growth of mutants is shown. Toxicity was assayed by growing cultures under permissive conditions (in LB-MES [pH 6.0]), followed by spotting 10-fold dilutions on LB-MOPS (pH 7.7) agar containing 86 mM Na⁺.

Fig. 4.

Phenotypic analysis of tobramycin-sensitive mutants. Antibiotic sensitivities (A) and pH and NaCl sensitivities (B) of mutants were characterized using phenotype microarray analysis. Each panel shows the tetrazolium reduction during 48 h, a measure of respiration. Tetrazolium reduction by each mutant (green) is compared to that of a control strain (red), with overlapping regions in yellow. Tested mutants were compared to the PA3303 insertion control, with the exception of ΔftsH, which was compared to MPAO1. For antibiotics, growth across four progressively higher concentrations is presented from left to right. The following mutants were tested (see Table S1 in the supplemental material): AHP36, AHP126, AHP83, AHP124, AHP125, AHP147, AHP146, AHP145, AHP127, AHP241, and SLP1742.

Fig. 5.

Functions responsible for intrinsic aminoglycoside resistance. Relationships among the genes contributing to tobramycin resistance are presented. For clarity, the MexXY-OprM efflux pump, MexF-PhoA, and other resistance factors are not included.