Sensing and adaptation to low pH mediated by inducible amino acid decarboxylases in Salmonella

Abstract

During the course of infection, Salmonella enterica serovar Typhimurium must successively survive the harsh acid stress of the stomach and multiply into a mild acidic compartment within macrophages. Inducible amino acid decarboxylases are known to promote adaptation to acidic environments. Three low pH inducible amino acid decarboxylases were annotated in the genome of S. Typhimurium, AdiA, CadA and SpeF, which are specific for arginine, lysine and ornithine, respectively. In this study, we characterized and compared the contributions of those enzymes in response to acidic challenges. Individual mutants as well as a strain deleted for the three genes were tested for their ability (i) to survive an extreme acid shock, (ii) to grow at mild acidic pH and (iii) to infect the mouse animal model. We showed that the lysine decarboxylase CadA had the broadest range of activity since it both had the capacity to promote survival at pH 2.3 and growth at pH 4.5. The arginine decarboxylase AdiA was the most performant in protecting S. Typhimurium from a shock at pH 2.3 and the ornithine decarboxylase SpeF conferred the best growth advantage under anaerobiosis conditions at pH 4.5. We developed a GFP-based gene reporter to monitor the pH of the environment as perceived by S. Typhimurium. Results showed that activities of the lysine and ornithine decarboxylases at mild acidic pH did modify the local surrounding of S. Typhimurium both in culture medium and in macrophages. Finally, we tested the contribution of decarboxylases to virulence and found that these enzymes were dispensable for S. Typhimurium virulence during systemic infection. In the light of this result, we examined the genomes of Salmonella spp. normally responsible of systemic infection and observed that the genes encoding these enzymes were not well conserved, supporting the idea that these enzymes may be not required during systemic infection.

Figure 1. Growth of WT and ΔadiAΔcadAΔspeF at moderate acidic pH.

Bacteria grown overnight in LBG pH 7 were washed and diluted to OD₆₀₀ = 0.03 in M9 medium complemented with 0.1% casamino acids, 0.2% glucose and adjusted to pH 4.5 with HCl. When indicated 5 mM L-ornithine, L-lysine or L-arginine were added to the medium. Cultures were performed in aerobic (A-B) or anoxic (C-D) conditions and monitored by following optical density at 600 nm. Typical growth curves representative of several experiments are shown. WT is the wild-type strain; ΔadiAΔcadAΔspeF is the strain deleted for the three genes adiA, cadA, and speF (strain n°197).

Figure 2. Growth of the ΔcadA and ΔspeF mutants in anoxic conditions at moderate acidic pH.

Bacteria grown overnight in LBG pH 7 were washed and diluted to OD₆₀₀ = 0.03 in M9 medium complemented with 0.1% casamino acids, 0.2% glucose and adjusted to pH 4.5 with HCl. When indicated 5 mM L-ornithine and/or 5 mM L-lysine were added to the medium. Cultures were performed in anoxic conditions and monitored by following optical density at 600 nm.

Figure 3. Complementation by pcadA and analysis of ΔadiAΔspeF during growth at moderate acidic pH.

Bacteria grown overnight in LBG pH 7 were washed and diluted to OD₆₀₀ = 0.03 in M9 medium complemented with 5 mM L-lysine, 0.1% casamino acids, 0.2% glucose and adjusted to pH 4.5 with HCl. pACYC177 is a low-copy cloning vector and pcadA is pACYC177 in which the cadA gene has been introduced under control of its natural promoter. Cultures were either performed in anoxic (A) or aerobic (B) conditions and monitored by following optical density at 600 nm. Panels C and D show growth of the WT and ΔadiAΔspeF strains in the same M9 medium complemented or not with 5 mM lysine, at pH 4.5 in anoxic (C) and aerobic (D) conditions.

Figure 4. Characterization of the transcriptional fusion Pasr::gfp in acidified M9 medium.

Fluorescence produced by WT pPasr::gfp was monitored with a fluorometer and expressed proportionally to the bacterial population (GFP/OD₆₀₀). A. WT pPasr::gfp was grown overnight in M9 medium complemented with 0.1% casamino acids, 0.2% glucose pH 7.2, diluted 1/50 and grown in the same medium adjusted to the desired pH with HCl. B. Bacteria were prepared as in A and grown in M9 medium adjusted to pH 4.5 or left at pH 7.2 with addition of different stresses: amino acids, magnesium or iron starvations (a.a._-less, Mg²⁺ _-less and Fe²⁺ _-less, respectively), oxidative stress (hydrogen peroxide : H₂O₂) and antimicrobial peptide (polymixin B : PB). C. WT pPasr::gfp and ΔadiAΔcadAΔspeF pPasr::gfp were grown overnight in the same medium as in A, diluted 1/50 and grown in the same medium containing, when indicated, 5 mM L-lysine and 5 mM L-ornithine and adjusted to pH 4.5. Cultures were performed in anoxic conditions.

Figure 5. Response of the Pasr::gfp transcriptional fusion during macrophage infection.

WT pPasr::gfp was used to infect RAW264.7 macrophages. At various times post-infection (pi) host cells were lysed, bacteria were fixed with paraformaldehyde and collected. Fluorescence produced by intracellular WT pPasr::gfp 1, 5 and 8 hours post-infection was analyzed by flow cytometry (A). The same analysis was performed when host cells were treated with 100 nM bafilomycin A1, which inhibits SCV acidification (B). Typical graphs representative of several experiments are shown.

Figure 6. Delay in SCV acidification when decarboxylase activities are favoured.

WT pPasr::gfp and ΔadiAΔcadAΔspeF pPasr::gfp were grown for 5 hours in anoxia in LBG and used to infect RAW264.7 macrophages. When indicated L-lysine and L-ornithine, 20 mM each, were added in the cell culture medium 3 hours before infection and throughout the experiment. At 1h30 (A, C), 4h (B, C) and 7h30 (C) post-infection, macrophages were lysed, bacteria were fixed with paraformaldehyde and collected. Fluorescence produced by WT pPasr::gfp and ΔadiAΔcadAΔspeF pPasr::gfp was analyzed by flow cytometry. In C, curves were produced using the mean FL1-H corresponding to the mean fluorescence intensity of the bacterial population for one time point.

Figure 7. Competitive index in mice between WT and mutants of the inducible amino acid decarboxylases.

Mice were orally inoculated with an equal mix of wild-type and mutant strains, spleens were analyzed 5 days post-inoculation. The CI was calculated as the output ratio of mutant to wild-type bacteria divided by the input ratio. Open symbols illustrate individual CI values, horizontal bars indicate means of CI values and error bars represent the standard error of the mean. Abbreviation ns means not significantly different from wild-type, which occurs when p-value (p) is superior to 0.05. The ΔadiAΔcadAΔspeF strain is the kanamycin resistant version of ΔadiAΔcadAΔspeF (strain n° 199).