Genome-wide transposon mutagenesis of Proteus mirabilis: Essential genes, fitness factors for catheter-associated urinary tract infection, and the impact of polymicrobial infection on fitness requirements

Abstract

The Gram-negative bacterium Proteus mirabilis is a leading cause of catheter-associated urinary tract infections (CAUTIs), which are often polymicrobial. Numerous prior studies have uncovered virulence factors for P. mirabilis pathogenicity in a murine model of ascending UTI, but little is known concerning pathogenesis during CAUTI or polymicrobial infection. In this study, we utilized five pools of 10,000 transposon mutants each and transposon insertion-site sequencing (Tn-Seq) to identify the full arsenal of P. mirabilis HI4320 fitness factors for single-species versus polymicrobial CAUTI with Providencia stuartii BE2467. 436 genes in the input pools lacked transposon insertions and were therefore concluded to be essential for P. mirabilis growth in rich medium. 629 genes were identified as P. mirabilis fitness factors during single-species CAUTI. Tn-Seq from coinfection with P. stuartii revealed 217/629 (35%) of the same genes as identified by single-species Tn-Seq, and 1353 additional factors that specifically contribute to colonization during coinfection. Mutants were constructed in eight genes of interest to validate the initial screen: 7/8 (88%) mutants exhibited the expected phenotypes for single-species CAUTI, and 3/3 (100%) validated the expected phenotypes for polymicrobial CAUTI. This approach provided validation of numerous previously described P. mirabilis fitness determinants from an ascending model of UTI, the discovery of novel fitness determinants specifically for CAUTI, and a stringent assessment of how polymicrobial infection influences fitness requirements. For instance, we describe a requirement for branched-chain amino acid biosynthesis by P. mirabilis during coinfection due to high-affinity import of leucine by P. stuartii. Further investigation of genes and pathways that provide a competitive advantage during both single-species and polymicrobial CAUTI will likely provide robust targets for therapeutic intervention to reduce P. mirabilis CAUTI incidence and severity.

Fig 1. Saturation of the Proteus mirabilis HI4320 genome and plasmid pHI4320 with transposon insertions.

Chromosomal and plasmid maps (not to scale) indicating the location of all transposon insertions contained in all five of the input pools. On average, the P. mirabilis transposon pools contained an insertion every 50.9 bp, with ~21,694 open reading frame insertions per pool. Regions without insertions represent estimated essential genes and intergenic regions. Each line represents a single insertion site from the input samples, and the length of the line represents the log₁₀ number of reads recovered from each insertion site. The color of the line indicates the number of pools in which each insertion site was identified: all five pools (green), four pools (blue), three pools (purple), two pools (red), one pool (black).

Fig 2. Functional categories of P. mirabilis HI4320 estimated essential genes.

A Bayesian mixture model was used to identify genes estimated to be essential for P. mirabilis growth in LB medium based on absence or underrepresentation of transposon insertions in these genes from all five input pools (see Materials and methods). (A) The number of annotated genes encoded by HI4320 pertaining to each Cluster of Orthologous Groups of proteins (COG) is indicated by the light gray bars. Within each COG, the genes estimated to be essential for growth in LB broth are indicated by the dark gray bars, and the number of these estimated essential genes present in the Database of Essential Genes (DEG) is indicated by crosshatching. (B) The percentage of estimated essential genes belonging to each COG are displayed in descending order.

Fig 3. Conceptual model of single-species and polymicrobial CAUTI Tn-Seq.

For each of five transposon mutant library pools, mice were infected as follows: 1) 5–10 CBA/J mice were transurethrally inoculated with 1x10⁵ CFU of the transposon library for single-species infection, and 2) 5–10 CBA/J mice were inoculated with 1x10⁵ CFU of a 1:1 mixture of the transposon library and wild-type P. stuartii BE2467 (purple) for coinfection. Thus, for each input pool, the single-species infections and coinfections were conducted in parallel to utilize the same input inoculum. Input and output samples were enriched for transposon-containing sequences and subjected to next generation Illumina sequencing of the transposon-chromosome junctions. The resulting reads were mapped to the P. mirabilis genome, and the abundance of reads at each insertion site from all output samples were compared to the input samples to determine a fold change for each gene. The gene in yellow represents a candidate P. mirabilis fitness factor for single-species CAUTI that is even more important during coinfection; the gene in blue represents a P. mirabilis fitness factor for single-species CAUTI that is no longer important during coinfection; the gene in red represents a factor that does not contribute to P. mirabilis CAUTI and was therefore recovered at a similar density from the infection output pools as the input pools.

Fig 4. Colonization by P. mirabilis transposon mutants during single-species and polymicrobial CAUTI.

For each transposon mutant library pool, 5–10 CBA/J mice were transurethrally inoculated with 1x10⁵ CFU of the transposon library for single-species infection (A), and 5–10 CBA/J mice were inoculated with 1x10⁵ CFU of a 1:1 mixture of the transposon library and wild-type P. stuartii BE2467 for coinfection (B and C). In all cases, a 4 mm segment of catheter tubing was retained in the bladder for the duration of the study. Mice were sacrificed 4 days post-inoculation, and the bladder and kidneys were homogenized, and an aliquot was plated onto LB agar to determine bacterial burden. The remaining homogenate was fully plated for isolation of bacterial genomic DNA and sequencing. (A and B) Each symbol represents total CFU/gram of tissue from an individual mouse during single-species infection (A) or coinfection (B), and error bars indicate the median. (C) The majority of coinfected mice were highly colonized by both bacterial species. Blue circles represent P. mirabilis CFU/gram of tissue and red triangles represent P. stuartii CFU/gram of tissue, with values from a single mouse connected by a black line. Dashed lines indicate limit of detection.

Fig 5. Functional categories of P. mirabilis fitness factors for single-species CAUTI.

(A) The percentage of 203 candidate P. mirabilis fitness factors required for colonization of both the catheterized bladder and kidneys during single-species CAUTI belonging to each COG are displayed in descending order. (B) COG categories represented by 64 candidate fitness factors for bladder colonization that did not contribute to kidney colonization. (C) COG categories represented by 127 candidate fitness factors for kidney colonization that did not contribute to bladder colonization.

Fig 6. Growth of wild-type P. mirabilis and mutants in LB broth and minimal medium.

Growth of P. mirabilis HI4320 and mutants was measured in LB broth (A and B) and PMSM minimal medium (C, D, and E) with the following supplements: (A) none, (B) polymyxin B, (C) 1 mM l-glutamine, (D) 10 mM total of the listed BCAAs, and (E) nickel sulfate. Graphs are representative of at least three independent experiments. Error bars represent mean ± standard deviation (SD) from at least five technical replicates. Differences in growth between strains and treatment conditions were determined to be significant by two-way ANOVA.

Fig 7. Urease activity and motility of wild-type P. mirabilis and mutants.

(A) P. mirabilis and mutants were cultured in filter-sterilized human urine to measure urease activity, expressed as the mean change in optical density per minute (mOD/min) of the indicator dye phenol red during a 5-minute kinetic read at each indicated timepoint. Graph is representative of at least three independent experiments. Error bars represent mean ± SD from at least three technical replicates. ***P<0.001 by two-way ANOVA. (B) Swimming motility diameter in Mot agar, compiled from three independent experiments with three technical replicates each. Error bars represent mean ± SD. *P<0.05, ***P<0.001 compared to wild-type by Student’s t-test. (C) Diameter of the 1^st (R1), 2^nd (R2), and 3^rd (R3) swarm rings to the consolidation zone and total swarm diameter for P. mirabilis and mutants compiled from three independent experiments with at least two technical replicates each. Dashed lines indicate average swarm ring diameter for wild-type P. mirabilis. Error bars represent mean ± SD. *P<0.05, **P<0.01, and ***P<0.001 compared to wild-type P. mirabilis by two-way ANOVA with post-hoc test for significance.

Fig 8. Validation of candidate P. mirabilis fitness factors for single-species CAUTI.

CBA/J mice were transurethrally inoculated with 1x10⁵ CFU of a 1:1 mixture of wild-type P. mirabilis and an isogenic mutant. In all cases, a 4 mm segment of catheter tubing was retained in the bladder for the duration of the study. Urine was collected and mice were sacrificed 4 days post-inoculation, and the catheterized bladder, kidneys, and spleen were homogenized and plated onto LB agar with and without kanamycin to determine bacterial burden of wild-type P. mirabilis and the mutant (A-H). Each data point represents the Log₁₀ CFUs recovered from an individual mouse, and gray bars represent the median. A competitive index was calculated for each mutant on a per-mouse basis, for organs in which the mutant and wild-type were above the limit of detection, using the ratio of mutant to wild-type in each organ divided by the ratio of mutant to wild-type from the inoculum (I-P, see Materials and methods). (A and I) lon, (B and J) argR, (C and K) hslU, (D and L) arnA, (E and M) PMI1518, (F and N) glnA, (G and O) pldA, and (H and P) ilvD. Solid lines represent the median. Dashed lines indicate a competitive index of 1, or a 1:1 ratio of mutant to wild-type. *P<0.05 and **P<0.01 by the Wilcoxon signed rank test.

Fig 9. In vitro co-challenge of P. mirabilis mutants in human urine.

Filter-sterilized pooled human urine from healthy donors was inoculated with a 1:1 mixture of wild-type P. mirabilis and the following mutants that exhibited significant fitness defects in vivo: (A and B) lon, (C and D) argR, (E and F) hslU, (G and H) arnA, (I and J) PMI1518, and (K and L) glnA. Cultures were incubated at 37°C for 5 hours, and sampled hourly for determination of CFUs (A, C, E, G, I, and K). Error bars represent mean and error for three independent replicates. No differences in growth between mutants and wild-type were detected by two-way ANOVA with post-hoc test for significance. A competitive index was calculated for each mutant at each hourly timepoint using the ratio of mutant to wild-type at the time of inoculation (B, D, F, H, J, and L). None of the mutants exhibited a significant fitness defect or advantage during growth in urine by the Wilxocon signed rank test.

Fig 10. Functional categories of P. mirabilis fitness factors for polymicrobial CAUTI.

(A) The percentage of 717 candidate P. mirabilis fitness factors required for both bladder and kidney colonization during single-species CAUTI belonging to each COG are displayed in same the order as in Fig 5 for comparison. (B) COG categories of 45 candidate fitness factors for bladder colonization that did not contribute to kidney colonization. (C) COG categories of 157 candidate fitness factors for kidney colonization that did not contribute to bladder colonization.

Fig 11. Validation of candidate P. mirabilis fitness factors for polymicrobial CAUTI.

CBA/J mice were transurethrally inoculated with 1x10⁵ CFU of the following mixture: 5x10⁴ CFUs of a 1:1 mixture of the mutant and wild-type P. mirabilis, and 5x10⁴ CFUs of wild-type P. stuartii. In all cases, a 4 mm segment of catheter tubing was retained in the bladder for the duration of the study. Urine was collected and mice were sacrificed 4 days post-inoculation, and the bladder, kidneys, and spleen were homogenized and plated onto LB agar with and without kanamycin to determine bacterial burden of P. stuartii, wild-type P. mirabilis, and the mutant (A-C). Each data point represents the Log₁₀ CFUs recovered from an individual mouse, and gray bars represent the median. A competitive index was calculated for each P. mirabilis mutant on a per-mouse basis using the ratio of mutant to wild-type in each organ divided by the ratio of mutant to wild-type from the inoculum to determine if the presence of P. stuartii resulted in the P. mirabilis mutant being significantly outcompeted by its parental wild-type strain (D-F). (A and D) the PMI1518 mutant vs wild-type P. mirabilis during coinfection with P. stuartii, (B and E) lon, and (C and F) ilvD. Solid lines represent the median. Dashed lines indicate a competitive index of 1, or a 1:1 ratio of mutant to wild-type. *P<0.05, **P<0.01, and ***P<0.001 by the Wilcoxon signed rank test.

Fig 12. BCAA import and biosynthesis contribute to P. stuartii and P. mirabilis fitness during coinfection.

CBA/J mice were transurethrally inoculated with 1x10⁵ CFU and a 4 mm segment of catheter tubing was retained in the bladder for the duration of the study. Urine was collected and mice were sacrificed 4 days post-inoculation, and the bladder, kidneys, and spleen were homogenized and plated onto LB agar with and without kanamycin to determine bacterial burden of P. stuartii, P. mirabilis, and their respective mutants (A-D), competitive indices were calculated for each infection (E-H). Mice were inoculated with the following mixtures: (A and E) 1:1 mixture of the P. stuartii livK mutant and wild-type P. stuartii, (B and F) 5x10⁴ CFU of a 1:1 mixture of the P. stuartii livK mutant and wild-type P. stuartii and 5x10⁴ CFU of wild-type P. mirabilis, (C and G) 5x10⁴ CFUs of a 1:1 mixture of the P. mirabilis ilvD mutant and wild-type P. mirabilis and 5x10⁴ CFUs of wild-type P. stuartii, and (D and H) 5x10⁴ CFUs of a 1:1 mixture of the P. mirabilis ilvD mutant and wild-type P. mirabilis and 5x10⁴ CFUs of the P. stuartii livK mutant. A competitive index was calculated for the P. stuartii livK mutant during single-species co-challenge (E) and co-challenge during coinfection with wild-type P. mirabilis (F), and a competitive index was calculated for the P. mirabilis ilvD mutant during coinfection with wild-type P. stuartii (G) or with the P. stuartii livK mutant (H). Error bars represent the median. Dashed lines indicate a competitive index of 1, or a 1:1 ratio of mutant to wild-type. *P<0.05 by the Wilcoxon signed rank test.