The Basis for Natural Multiresistance to Phage in Pseudomonas aeruginosa

Abstract

Pseudomonas aeruginosa is responsible for long-term infections and is particularly resistant to treatments when hiding inside the extracellular matrix or biofilms. Phage therapy might represent an alternative to antibiotic treatment, but up to 10% of clinical strains appear to resist multiple phages. We investigated the characteristics of P. aeruginosa clinical strains naturally resistant to phages and compared them to highly susceptible strains. The phage-resistant strains were defective in lipopolysaccharide (LPS) biosynthesis, were nonmotile and displayed an important degree of autolysis, releasing phages and pyocins. Complete genome sequencing of three resistant strains showed the existence of a large accessory genome made of multiple insertion elements, genomic islands, pyocins and prophages, including two phages performing lateral transduction. Mutations were found in genes responsible for the synthesis of LPS and/or type IV pilus, the major receptors for most phages. CRISPR-Cas systems appeared to be absent or inactive in phage-resistant strains, confirming that they do not play a role in the resistance to lytic phages but control the insertion of exogenous sequences. We show that, despite their apparent weakness, the multiphage-resistant strains described in this study displayed selective advantages through the possession of various functions, including weapons to eliminate other strains of the same or closely related species.

Keywords: CRISPR-Cas systems; Pseudomonas aeruginosa; bacteriophage; coevolution; genome sequence; lateral transduction; mechanisms of resistance; mobile genetic elements; phage therapy; prophage induction.

Figure 1

SDS-PAGE of the lipopolysaccharide (LPS) fraction. SDS-PAGE of the LPS fraction. The position of the core, the A-band and the B-band of the O-antigen were identified according to the work of Islam et al. [50]. In order to clearly detect the different molecular species present in the LPS profiles, we had to use different amounts of samples when looking at the long O-chains (left panel) and at the shorter core LPS (right panel). We used 12.5 µg of suspended lyophilized bacteria for the gel on the left panel and 2.5 µg for the gel on the right panel. The resistant strains are shown with red letters.

Figure 2

Phenotype of the strains. (A) Twitching motility is expressed as the diameter (cm) of the growth zone at the bottom of the agar plate. The standard deviation is the result of three independent assays. (B) Biofilm formation assay. The amount of bacteria bound to the wells is evaluated by measuring the OD₆₀₀ of crystal violet resuspended in ethanol. The standard deviation is the result of three independent assays.

Figure 3

Electron microscopy analysis of particles induced by mitomycin C. (A) PAC7-25 where the white arrow points to the phage tail, (B) PcyII-29 and (C) PcyII-40 where the white arrows point to vesicles, and the dark arrow points to a group of two attached ~50-nm tubes. Pyocins are indicated with a P and bacteriophages with a B. The black bar represents 100 nm.

Figure 4

Mapping of sequencing reads from MitC-induced virions onto bacterial genomes. About one million reads were used in Geneious PcyII-29 and PcyII-40. The arrow points to reads mapping on chromosomal DNA.

Figure 5

Mapping of the MitC-induced phages’ reads from PcyII-40 onto the PAO1_Or chromosome. A portion of the chromosome corresponding to the insertion site of phage PCYII-40_MGE8 is shown. The larger peak of reads is at the left side of the attB site and covers about 30 kb. Three additional peaks covering 46kb are observed.