Massively parallel mutant selection identifies genetic determinants of Pseudomonas aeruginosa colonization of Drosophila melanogaster

Abstract

Pseudomonas aeruginosa is recognized for its ability to colonize diverse habitats and cause disease in a variety of hosts, including plants, invertebrates, and mammals. Understanding how this bacterium is able to occupy wide-ranging niches is important for deciphering its ecology. We used transposon sequencing [Tn-Seq, also known as insertion sequencing (INSeq)] to identify genes in P. aeruginosa that contribute to fitness during the colonization of Drosophila melanogaster. Our results reveal a suite of critical factors, including those that contribute to polysaccharide production, DNA repair, metabolism, and respiration. Comparison of candidate genes with fitness determinants discovered in previous studies on P. aeruginosa identified several genes required for colonization and virulence determinants that are conserved across hosts and tissues. This analysis provides evidence for both the conservation of function of several genes across systems, as well as host-specific functions. These findings, which represent the first use of transposon sequencing of a gut pathogen in Drosophila, demonstrate the power of Tn-Seq in the fly model system and advance the existing knowledge of intestinal pathogenesis by D. melanogaster, revealing bacterial colonization determinants that contribute to a comprehensive portrait of P. aeruginosa lifestyles across habitats.IMPORTANCEDrosophila melanogaster is a powerful model for understanding host-pathogen interactions. Research with this system has yielded notable insights into mechanisms of host immunity and defense, many of which emerged from the analysis of bacterial mutants defective for well-characterized virulence factors. These foundational studies-and advances in high-throughput sequencing of transposon mutants-support unbiased screens of bacterial mutants in the fly. To investigate mechanisms of host-pathogen interplay and exploit the tractability of this model host, we used a high-throughput, genome-wide mutant analysis to find genes that enable the pathogen P. aeruginosa to colonize the fly. Our analysis reveals critical mediators of P. aeruginosa establishment in its host, some of which are required across fly and mouse systems. These findings demonstrate the utility of massively parallel mutant analysis and provide a platform for aligning the fly toolkit with comprehensive bacterial genomics.

Keywords: TnSeq; host–microbe interactions; host–pathogen interactions; oral infections.

Fig 1

Using the Tn-Seq platform to identify colonization factors of P. aeruginosa in the fly. (A) A highly saturated P. aeruginosa transposon library was constructed, prepared for Tn-Seq, and characterized (“input”). Approximately 62,000 insertions—47,000 of which were unique (37,000 in ORFs and 10,000 in intergenic regions)—were mapped to the P. aeruginosa PAO1 genome. The color intensity increases with the increasing number of insertions (white <10 insertions; pink = 11–100 insertions; red = 101–1,000 insertions; sites with >1,000 insertions are black). GC content is also shown (dark blue: greater than average; light blue: less than average; GC content = 66.56%). (B) 3–7-day-old female Canton-S flies housed in capillary feeders (CAFEs; a cartoon is shown above) were fed a population of PAO1 transposon mutants (input) for 24 h and then were switched to a sucrose–LB broth suspension for an additional 48 h (n = 250 flies). After the input was administered to flies; the fly homogenate was cultured; and P. aeruginosa mutants were recovered, prepared for Tn-Seq, and characterized (“output”). Genes underrepresented after passage through the fly were considered “putative colonization factors.” Genes with no insertions in the input were not analyzed in the output. Flies were sampled 72 h post-challenge. Six independent replicates were performed. (C and D) 3–7-day-old female Canton-S flies were fed on a 10⁵ CFU/mL suspension in 5% sucrose for 24 h. Then, they were transferred to 5% sucrose in LB medium and maintained in the sucrose solution for the duration of the experiment. Flies were sampled daily for the 5-day period. For the colony-forming units (CFUs) per fly (C), means and SEM are shown. The limit of detection is at the axis (100 CFUs per fly). n = 8 flies per group per day sampled. For fly survival (D), n = 20 flies.

Fig 2

Comparison of P. aeruginosa fitness determinants in vitro and in vivo. (A) After characterizing a mock output population that was subjected to experimental conditions, but not administered to flies, we compared genes depleted in that group (n = 379) to the output population recovered from flies (n = 372). Eighty-five genes were depleted only in vitro; 78 genes were depleted only in vivo. Genes that were negatively selected in vivo, but not in vitro, were termed putative “colonization-specific factors.” Nearly 300 genes were negatively selected both while the library was administered to flies (in vitro) and during passage through the fly (in vivo). (B) Genes depleted specifically during administration of the library tended to be categorized as contributing to the secondary structure (COG category Q), nucleotide metabolism and transport (F), lipid metabolism (I), and energy production and conversion (C), whereas genes depleted specifically in the fly were tended to be categorized as contributing to transcription, translation, signal transduction, replication and repair, cell motility, and amino acid metabolism and transport. COG category B (chromatin structure and dynamics), W (extracellular structures), and Z (cytoskeleton) and mobile elements were not represented among these genes specifically depleted in either condition. COG = Clusters of Orthologous Groups. Full lists of genes can be found in Tables S2 to S5 in reference .

Fig 3

Comparison of genetic determinants of P. aeruginosa colonization in the fly and mouse from previous studies. (A) Alginate and psl polysaccharides, purines, pyrimidines, amino acids, cofactors, and respiration genes are critical for establishment of PAO1 and P. aeruginosa PA14 in invertebrate and vertebrate hosts and at different body sites (14, 34). (B) Mutants in aceA and pilM were positively selected across systems, in both PAO1 and PA14. Most of the genes encoding components of flagella and pili were positively selected in the mouse but not in the fly. Insertions in the mexEF-oprN operon, exsB, pagL, and bifA, among others, were overrepresented in the fly.

Fig 4

Validation of fitness determinants identified by Tn-Seq. A 1:1 ratio of wild-type PAO1 and the indicated transposon mutant was administered to flies for 24 h. Flies were then given 5% sucrose for 48 h and homogenized at 72 h post-challenge. Bacteria were cultured on selective and non-selective media. Each symbol represents the CFU recovered from an individual fly, where n = 8 flies per group and one representative replicate is shown. Competitive Index = Log10(CFU mutant/ CFU wild type), where a competitive index of 0 indicates equal competitive fitness.