Novel Aminoglycoside-Tolerant Phoenix Colony Variants of Pseudomonas aeruginosa

Bacterial strain and culture conditions.The bioluminescent strain P. aeruginosa Xen41 (Xenogen Corp., USA) and its parent P. aeruginosa PAO1 were used in this study. Glycerol stock cultures were stored at –80°C and streaked onto fresh Luria-Bertani (LB) agar plates that were incubated for 24 h. Isolated colonies from the LB agar plates were transferred to 20 ml of LB broth and incubated overnight on an incubator shaker set at a temperature of 37°C and at a speed of 200 rpm.

Preparing lawn biofilms of P. aeruginosa.Spreading the overnight culture on LB agar generated lawn biofilms of P. aeruginosa Xen41 and PAO1. Briefly, the overnight P. aeruginosa Xen41 or PAO1 culture grown in LB broth was diluted to an optical density at 600 nm (OD₆₀₀) of 0.1. Then, 100 μl of the diluted culture was spread onto a 100-mm diameter polystyrene petri plate (Fisher Scientific, USA) containing LB agar. The petri plates were incubated at 37°C for 24 h to develop a lawn biofilm of P. aeruginosa. Additionally, the CFU/cm² of the lawns was measured to be used as a reference value. In short, a biofilm lawn was generated, and a 1 cm by 1 cm area was marked and isolated using a sterile plastic loop in 1 ml of sterile phosphate-buffered saline (PBS). A dilution series was generated and plated onto LB agar by dripping 5 μl of each dilution onto the agar surface and allowing it to dry. The plates were then incubated at 37°C with 5% CO₂ overnight. Colonies were then counted, and the CFU/cm² of the lawn was calculated to be 5 ± 2 · 10⁹ CFU/cm².

Preparation of tobramycin-containing calcium sulfate beads.All experiments conducted required the preparation of CaSO₄ (Sigma-Aldrich) beads containing tobramycin. We used 240 mg of tobramycin (Sigma-Aldrich) per 20 g of CaSO₄, a ratio commonly used by orthopedic surgeons when mixing antibiotics into pharmaceutical grade CaSO₄ bone void filler for local release at the surgical site in PJIs (32). Once the tobramycin and CaSO₄ were mixed together, sterile water was added and mixed for approximately 1 min until a thick paste was produced. This paste was spread into silicone molds (Biocomposites Ltd.) to form hemispherical beads of 4.5 mm diameter and allowed to dry overnight.

Exposure of lawn biofilms to antibiotic-loaded beads.After generation of a 24-h lawn biofilm, a tobramycin-loaded bead was placed in the center of the plate using sterile forceps. The forceps were used to also push the bead into the agar. Once the bead was placed, the plate was incubated at 37°C with 5% CO₂ in a humidified incubator (Heracell 150i; Thermo Scientific) for 3 days and checked daily for the appearance and spread of a ZOC as well as any colonies appearing in this zone. In vitro imaging system (IVIS) images were taken daily to monitor bioluminescence, which is an indicator of metabolic activity within P. aeruginosa Xen41. Photographic images were also taken at the same time. After 3 days, IVIS and photographic images were taken again, and variant colonies in the zone of lawn clearing were manually counted.

Bioluminescence imaging.Bioluminescence imaging was performed using IVIS imaging. Thirty-second exposures were obtained for each plate imaged. A pseudo-color heatmap was applied where red indicates high light intensity, blue indicates low levels of light intensity, and black indicates no light present. While a black color indicates a lack of activity, it cannot be used to rule out cell viability.

Determination of sensitive and resistant strains using replica plating.To characterize the colonies in the cleared zone and to screen for persisters, we used replica plating (33). A sterile cotton velveteen square (150 by 150 mm) was aseptically draped over a polyvinyl chloride (PVC) replica plater and locked in place with an aluminum ring. The tobramycin-loaded bead was removed from the center of the plate after 5 days of incubation using a sterile plastic loop. The plate was then marked to indicate the 12 o’clock position, gently placed onto the velveteen square, and tapped gently to ensure complete contact of the plate with the velveteen. The plate was then removed from the replica plater, and a fresh, sterile LB agar plate containing 5 μg/ml of tobramycin and marked at the 12 o’clock position was placed on the velveteen square. This plate was also tapped gently to ensure complete contact with the velveteen surface and then removed. Finally, a fresh, sterile LB agar plate marked at the 12 o’clock position was placed on the velveteen square and tapped gently to ensure complete surface contact. Both of the fresh replica plates were incubated for 24 h at 37°C in an incubator with 5% CO₂. After incubation, the pattern of the plates was compared, and colonies which appeared on both plates were deemed to be resistant mutants. Total colony counts were compared between the original plate and replica plates. The difference between the two can be explained by the presence of persister cells, and thus, persister cell calculations were completed in this manner. Colonies which appeared on the original plates but on neither of the replica plates were deemed to be VBNCs. An additional population of colonies was observed which appeared on the LB agar replica plate but not on the LB agar replica plate containing tobramycin. Because they do not fit into any of the other categories, we termed these phoenix colonies. All phoenix colonies were isolated and subjected to MIC testing for confirmation of susceptibility to tobramycin. In addition to replica plating, colonies were isolated in LB broth to confirm the lack of growth of VBNCs in the replica plating results. In short, variant colonies were generated as above and then isolated using sterile pipette tips and dipped into 200 μl of LB broth containing 5 μg/ml of tobramycin in a well of a 96-well plate (Corning, Sigma-Aldrich). The pipette tip was then dipped into the corresponding well of another 96-well plate containing 200 μl of LB broth. A total of 96 colonies were isolated from three different plates for a total of 288 colonies. The plates were then incubated for 96 h at 37°C with 5% CO₂. After incubation, the turbidity in the wells of each plate was compared. VBNCs were determined by lack of growth in the LB broth as well as the corresponding well of LB broth containing the tobramycin. Percentages for the phoenix and resistant strains were found to be consistent in the VBNC broth study compared to replica plating. The data presented in Table 1 were derived by combining both replica plating counts and VBNC broth data. Briefly, the VBNC percentage was taken directly from the VBNC broth data. This value was then subtracted from 100%, and the remaining percentage was used to calculate the phoenix, resistant, and persister cell proportions based on colony counts from replica plating.

Alternative antibiotic testing.In addition to tobramycin, additional clinically relevant antibiotics were tested for the emergence of phoenix colonies. Gentamicin, ciprofloxacin, and colistin are routinely used to treat P. aeruginosa infections (34, 35). Each of these antibiotics was examined in studies similar to those described above, including variant colony generation and replica plating. Phoenix colonies emerged within the ZOC of gentamicin (Fig. S1) but not in the ZOC of ciprofloxacin (Fig. S2) or colistin (Fig. S3), indicating that this may be an aminoglycoside-exclusive phenotype.

Comparison of the emergence of phoenix colonies by strain and antibiotic delivery method.In order to examine the emergence of phoenix colonies in different strains of Pseudomonas as well as to determine if CaSO₄ directly linked this emergence to antibiotic delivery, comparison studies were completed. Overnight cultures of P. aeruginosa Xen41 and PAO1 were generated as described above. Each culture was diluted to an OD₆₀₀ of 0.1, and 100 μl of the solution was spread onto sterile LB agar. Six plates were spread with Xen41, and six plates were spread with PAO1. All plates were incubated at 37°C with 5% CO₂ overnight. After 24 h, a sterile filter disk (7 mm, Remel, Thermo Scientific) was placed in the center of three of the Xen41 and three of the PAO1 plates, and 10 μl of a 10-mg/ml tobramycin solution was placed on the disk to match the tobramycin potency in a tobramycin-loaded bead. Tobramycin-loaded beads were placed in the center of each of the six remaining plates as described above. All plates were incubated at 37°C with 5% CO₂ for 5 days post-initial antibiotic exposure. Replica plating was performed, and colony counts were obtained for each plate and are represented by the mean ± standard deviation (SD).

Growth rate studies of isolates.To determine whether the phoenix colonies might have a growth defect and to rule out the presence of small colony variants (SCVs), we performed growth curves and measured colony diameter as a function of time on the plates. Variant colony generation was performed as described above, and after generation and characterization of the variant isolate MIC, growth curves were generated for each isolate along with wild-type P. aeruginosa Xen41 growth curves. In short, isolates were placed in 200 μl of LB broth in a well of a 96-well plate. The plates were then placed in a plate reader (SpectraMax i3x; Molecular Devices) and incubated statically at 37°C for 15 h. OD₆₀₀ readings were taken every 10 min, with 5 s of orbital shaking occurring before each read. After the growth curves were generated, growth curve data from phoenix colony replicates were averaged to determine the mean OD₆₀₀ at each time point. The same was done for resistant isolates and wild-type samples. The data were plotted as the mean ± SD. The maximum growth rate for each of these mean curves was also calculated from the maximum slope during the exponential phase and is presented as the mean ± SD. In situ growth on the plate was also measured for variant colonies. Variant colonies were generated as described above, and once colonies began to be visible, images were taken of the plates at various time points. After 5 days, the plates were replica plated, and each colony’s phenotype was determined. In addition to variant colonies, a dilution series of P. aeruginosa PAO1 was made and spread on several plates. These plates were then observed for the emergence of colonies, and once colonies were visible, images were taken of the plates at various time points. The images of all of the plates were then analyzed using image analysis software to measure the diameter of each colony. Measurements were grouped together based on variant colony phenotype, and the data were plotted as the mean ± SD.

Calculation of tobramycin concentration in agar plates and biofilm lawn.To measure the concentration of tobramycin in the agar and the biofilm as a function of time and radial distance from the tobramycin-loaded bead, we first generated a standard curve and then excised portions of both a biofilm lawn and agar exposed to a tobramycin-loaded bead in order to calculate the concentration of tobramycin in each by using a Kirby-Bauer-type assay. First, we generated calibration curves by diluting an overnight culture of P. aeruginosa PAO1 to an OD₆₀₀ of 0.1 and spreading the dilution onto sterile LB agar. A sterile filter disk was then placed in the center of the plate, and 10 μl of various tobramycin potencies (8 to 500 μg/ml) was placed onto the disk. The plate was incubated for 24 h at 37°C with 5% CO₂. After incubation, the zones of inhibition were measured and plotted to produce a standard curve (y = 5.2384ln[x] – 2.3996, R² = 0.9994).

Once the standard curve was generated, the tobramycin concentration in the agar at various radii and time points was measured. Tobramycin-loaded beads were placed into sterile LB agar plates. The plates were incubated at 37°C with 5% CO₂, and at various time points, plates were removed and marked with 5 mm by 5 mm squares from the edge of the bead to the edge of the plate. Each of these squares was excised using a razor blade and forceps. The excised plug was then melted at 80°C. An overnight culture of P. aeruginosa PAO1 was diluted to an OD₆₀₀ of 0.1 and spread onto sterile LB agar. A sterile filter disk was then placed in the center of the plate, and 10 μl of the melted plug was placed onto the disk. The plate was incubated for 24 h at 37°C with 5% CO₂. After incubation, the zone of inhibition was measured and compared to a standard curve for tobramycin in LB agar to determine the concentration in the melted plug. This process was repeated at the various time points, using different plates each time.

To quantify the amount of antibiotic that might have been portioned into the biofilm, we measured both the free antibiotic concentration in the lawn and the concentration trapped in the cells and biofilm extracellular polymeric substance (EPS). Variant colony generation was performed as described above. On day 4 post-tobramycin-loaded bead placement, the plates were removed from the incubator and marked with 5 mm by 5 mm squares from the edge of the bead to the edge of the plate. The biofilm lawn overlying each of the squares was isolated using a sterile plastic loop and placed in 1 ml of sterile double-distilled H₂O (ddH₂O). The sample was then vortexed for 10 s before being centrifuged at 21.1 × g for 2 min. The supernatant was collected, and the pellet was resuspended in 1 ml of sterile ddH₂O. The resuspended pellet was boiled at 100°C for 10 min to lyse the cells and denature proteins. An overnight culture of P. aeruginosa PAO1 was diluted to an OD₆₀₀ of 0.1 and spread onto sterile LB agar. A sterile filter disk was then placed in the center of the plate, and 10 μl of either the supernatant or boiled pellet was placed onto the disk. The plate was incubated for 24 h at 37°C with 5% CO₂. After incubation, the zone of inhibition was measured and compared to a standard curve for tobramycin in LB agar to determine the concentration in the respective components. Data are presented as the mean ± SD.

Population analysis profiling of background P. aeruginosa strains.To assess for the presence of heteroresistance (i.e., a small population of resistant cells that might go undetected by routine clinical antibiogram assays [20, 36]) in our background strains, we used population analysis profiling (PAP) (36). First, 20 ml of LB broth was inoculated with either P. aeruginosa PAO1 or P. aeruginosa Xen41. The cultures were incubated at 37°C with 200 rpm shaking for 24 h. After incubation, the cultures were diluted to an OD₆₀₀ of 0.1 and spread onto LB agar containing a 2-fold dilution series of tobramycin. The plates were then incubated at 37°C with 5% CO₂ for 24 h. After incubation, colony growth on each plate was analyzed. If heteroresistance is present in the population, colonies should appear on the dilution plate with an 8× higher concentration than the highest noninhibitory concentration. No heteroresistance was found in our background strains (data not shown).

Effect of efflux pump inhibitor on phoenix colony emergence.To assess the role of efflux pumps in the emergence of phoenix colonies, phe-arg-β-naphthylamide (PAβN), a broad-spectrum efflux pump inhibitor (37), was used. Lawn biofilms of P. aeruginosa PAO1 were generated as described above on LB agar containing 27 μM PAβN as well as LB agar without PAβN. Tobramycin-loaded bead placement and variant colony quantification were completed as described above. The mean ± SD of the colony counts was obtained and plotted.

Heritability study of phoenix colonies.To assess whether it is possible to enrich for the phoenix colony phenotype within a population, heritability studies on phoenix colony isolates were performed. Variant colonies were generated as described above. Variant colony plates were replica plated, and all phoenix colonies were isolated and used to inoculate 20 ml of LB broth. The broth cultures were incubated for 24 h at 37°C with 200 rpm shaking. After incubation, each of the cultures was diluted to an OD₆₀₀ of 0.1. Variant colonies for each isolate were again generated as before and replica plated. Isolation of phoenix colonies, overnight growth, and variant colony generation for each isolate were repeated. Data are presented as the mean ± SD.

Anaerobic chamber studies.A previous study showed the importance of alternative metabolic pathways in antimicrobial tolerance (25). In order to assess the effects of other metabolic pathways on the emergence of phoenix colonies, studies were performed in an anaerobic chamber to deprive the bacteria of oxygen. Variant colonies were generated as described above aside from incubation parameters on both LB agar plates and LB agar plates containing 100 mM KNO₃. The plates were incubated at 37°C in an anaerobic chamber at all times. Inside the chamber, the plates were also placed in a sealable bag containing a paper towel soaked in water to prevent dehydration of the agar. The plates were allowed to grow for 5 days and were removed from the chamber. Biofilm lawn growth was observed on the plate and imaged by IVIS, and the plates were immediately replica plated as described above onto both LB agar and LB agar containing 5 μg/ml of tobramycin. These plates were incubated at 37°C with 5% CO₂ in a standard aerobic incubator. After 24 h, growth on these plates was examined, and IVIS images were taken.

Eradication of planktonic tolerant, resistant, and phoenix phenotypes.The results from our replica plating studies indicate that there is a zone surrounding the antibiotic source in which everything, including persister cells, appears to be sterilized. To determine whether all cells truly could be killed with a high enough concentration of antibiotics, we performed a planktonic assay to look for the concentration at which bacterial eradication occurs. First, 100 ml of LB broth was used to prepare an overnight broth culture of P. aeruginosa PAO1. The culture was incubated at 37°C with 200 rpm of shaking for 24 h. Then, 5-ml aliquots of the overnight culture were placed into 15-ml conical tubes, and tobramycin was added to each tube in order to produce a triplicate dilution series of antibiotics at 400, 500, 600, 700, 800, 900, and 1,000 μg/ml.

Due to the very high concentrations of tobramycin in our studies, we were concerned that there may be an osmolarity effect causing the lysing of bacteria. To control for this, a separate triplicate was prepared with 0.385 mg/ml of dextrose to produce an equivocal maximum osmolarity solution. The new cultures were then incubated for 48 h at 37°C with 200 rpm of shaking. After incubation, 1 ml of each culture was centrifuged at maximum speed for 5 min. The supernatant was then removed, and the pellet was resuspended in 200 μl of sterile LB broth. The resuspension was used to prepare a 10-fold dilution series in a 96-well plate containing sterile LB broth. Then, 5-μl aliquots were dropped onto a sterile LB agar plate. The plates were incubated at 37°C with 5% CO₂ for 7 days in order to allow any dormant or slowly growing phenotypes to appear before being counted. The plates were then counted and CFU were calculated. Data are presented as the mean ± SD.

Statistical analysis.All experiments were performed in a minimum of triplicates. Analysis of variance (ANOVA) was completed using GraphPad Prism version 8.2.1 for all studies, with a P value of 0.05 being considered significant. Data represented in the graphs are plotted as the mean ± SD.