Abstract
The antibiotic defense substance allicin (diallylthiosulfinate) is produced by garlic (Allium sativum L.) after tissue damage, giving garlic its characteristic odor. Allicin is a redox-toxin that oxidizes thiols in glutathione and cellular proteins. A highly allicin-resistant Pseudomonas fluorescens strain (PfAR-1) was isolated from garlic, and genomic clones were shotgun electroporated into an allicin-susceptible P. syringae strain (Ps4612). Recipients showing allicin-resistance had all inherited a group of genes from one of three similar genomic islands (GI), that had been identified in an in silico analysis of the PfAR-1 genome. A core fragment of 8-10 congruent genes with redox-related functions, present in each GI, was shown to confer allicin-specific resistance to P. syringae, and even to an unrelated E. coli strain. Transposon mutagenesis and overexpression analyses revealed the contribution of individual candidate genes to allicin-resistance. Moreover, PfAR-1 was unusual in having 3 glutathione reductase (glr) genes, two copies in two of the GIs, but outside of the core group, and one copy in the PfAR-1 genome. Glr activity was approximately 2-fold higher in PfAR-1 than in related susceptible Pf0-1, with only a single glr gene. Moreover, an E. coli Δglr mutant showed increased susceptibility to allicin, which was complemented by PfAR-1 glr1. Taken together, our data support a multi-component resistance mechanism against allicin, achieved through horizontal gene transfer during coevolution, and allowing exploitation of the garlic ecological niche. GI regions syntenic with PfAR-1 GIs are present in other plant-associated bacterial species, perhaps suggesting a wider role in adaptation to plants per se.
