Refined methodology for quantifying Pseudomonas aeruginosa virulence using Galleria mellonella

ABSTRACT

Larvae of Galleria mellonella (the greater wax moth) are being increasingly used as a model to study microbial pathogenesis. In this model, bacterial virulence is typically measured by determining the 50% lethal dose (LD₅₀) of a bacterial strain or mutant. The use of G. mellonella to study Pseudomonas aeruginosa pathogenesis, however, is challenging because of the extreme sensitivity of larvae to this bacterium. For some P. aeruginosa strains, as few as 1–5 colony-forming units are sufficient to kill G. mellonella, which poses challenges for determining LD₅₀ values. For this reason, some groups have used time-to-death as a measure of P. aeruginosa virulence, but methodologies have not been standardized. We provide a detailed protocol for using the time at which 50% of larvae have died (LT₅₀) at a particular inoculum as a measure of P. aeruginosa virulence. We also describe a quality control metric for enhancing the reproducibility of LT₅₀ values. This approach provides an accurate and reproducible methodology for using G. mellonella larvae to measure and compare the virulence of P. aeruginosa strains.

IMPORTANCE

Pseudomonas aeruginosa is a significant cause of morbidity and mortality. The invertebrate Galleria mellonella is used as a model to determine the virulence of P. aeruginosa strains. We provide a protocol and analytical approach for using a time-to-death metric to accurately quantify the virulence of P. aeruginosa strains in G. mellonella larvae. This methodology, which has several advantages over 50% lethal dose approaches, is a useful resource for the study of P. aeruginosa pathogenicity.

INTRODUCTION

Galleria mellonella, also known as the greater wax moth, is found throughout the world. In nature, this organism has a parasitic relationship with honeybees, infiltrating hives and damaging colonies (1). Commercially, the larvae are reared as food for reptiles and as live bait for fishermen. In recent years, G. mellonella has been used extensively as a model system for the study of microbial pathogens (2). The larvae have several advantages over mice as a model system for the study of microbial pathogenesis (2). They are invertebrates and thus facilitate the goal of reducing the use of mammals in research (3). This leads to minimal regulatory burdens and allows for the use of death as an end point. They are inexpensive, easy to manipulate and maintain, and survive at 37°C, which approximates the internal human body temperature, an advantage when studying temperature-regulated virulence factors (4). Some components of the larval immune system, such as hemocytes and antimicrobial peptides, mimic those of the human innate immune system (5). Together, these features make G. mellonella an attractive model organism for studying the pathogenesis of certain bacteria.

Pseudomonas aeruginosa is a gram-negative, opportunistic pathogen that is a prominent cause of infections in hospitalized patients, those with underlying comorbidities, and individuals with cystic fibrosis (6). To facilitate infection, this pathogen uses a myriad of virulence determinants, including a type III secretion system, three type VI secretion systems, type IV pili, quorum sensing, and pyocyanin biosynthesis (6). Population genomics and systems biology approaches are emerging as powerful tools for exploring P. aeruginosa pathogenesis and treatment (7). However, these studies often require characterization of the virulence of a large number of P. aeruginosa strains in the context of a host, a need for which mice are less well suited than G. mellonella.

Most studies that use G. mellonella as a model for P. aeruginosa infection have measured 50% lethal dose (LD₅₀) values to quantify virulence (8–12). However, unlike many other bacterial pathogens, P. aeruginosa is extremely pathogenic to G. mellonella, making these virulence studies challenging. For example, whereas approximately 10⁵ colony-forming units (CFUs) of Klebsiella pneumoniae (13) and 10⁶ CFUs of Staphylococcus aureus (14) are required to kill G. mellonella, as few as 1–5 CFUs of many strains of P. aeruginosa are sufficient for this purpose (12, 15). For this reason, accurate LD₅₀ values are sometimes difficult to obtain. The second problem is that reproducible inoculation of such low numbers of bacteria is technically challenging and is frequently associated with large relative errors that may dramatically change reported survival values. Even when inoculums are accurately measured by plating following the experiment, it is difficult to reproducibly prepare the same inoculum size for distinct P. aeruginosa strains at the time of injection, making strain-to-strain comparisons of virulence problematic. Recently, some investigators have attempted to circumvent these difficulties by using time-to-death rather than lethal dose as a measure of P. aeruginosa virulence in G. mellonella (16). In this approach, the 50% lethal time (LT₅₀)—the time following inoculation at which 50% of larvae have succumbed from a specified dose of a microbe—is used to quantify virulence. However, standardization and general applicability of this methodology for comparing the virulence of P. aeruginosa strains have not been systematically investigated. Here, we provide a protocol for obtaining LT₅₀ values in G. mellonella to quantify P. aeruginosa virulence.

MATERIALS AND METHODS

Bacterial strains

P. aeruginosa bloodstream isolates PABL038, PABL051, PABL066, PABL083, PABL089, and PABL093 were originally obtained from the Clinical Microbiology Laboratory of Northwestern Memorial Hospital (Chicago, IL, USA) and stored at −80°C. These isolates and their corresponding LD₅₀ values in mice were previously published (17, 18).

The G. mellonella model

A detailed experimental protocol describing the G. mellonella methodology is provided in the supplemental Methods. Briefly, isolates were grown overnight at 37°C in lysogeny broth (LB) with shaking (250 RPM) and then sub-cultured in LB, pelleted, and resuspended in phosphate-buffered saline (PBS). Estimated doses for injection into larvae were obtained by diluting bacteria in PBS to an optical density at 600 nm (OD₆₀₀) of ~0.2 (the equivalent of ~5 × 10⁷ CFU/mL) and making appropriate dilutions with PBS to generate a range of bacterial doses (i.e., a range of suspensions of different CFU/mL) for injection into G. mellonella. Unless otherwise noted, larvae (Speedy Worm, Alexandria, MN, USA) were injected at the site of the final proleg with 10 µL of P. aeruginosa bacteria suspended in PBS using a Legato 100 Syringe Pump (KD Scientific, Holliston, MA, USA). Each inoculum for each P. aeruginosa isolate was injected into 10 larvae (i.e., 10 technical replicates). Injection with PBS alone was used as a negative control. Injected larvae were kept at 37°C in Petri dishes and monitored hourly for death, which was inferred by the absence of movement following gentle shaking. The number of dead larvae was recorded at each time point.

Analysis and statistics

For each dose and P. aeruginosa strain, an LT₅₀ value was calculated using the time of death of each of the 10 larvae by applying a custom R-Script utilizing the DRC package (https://github.com/ChrisAxline/Galleria_Code.git) (19). Because P. aeruginosa bacteria have been reported to multiply exponentially within G. mellonella (12, 15), we anticipated a logarithmic relationship between bacterial dose and time-to-death, so the natural logs of the P. aeruginosa doses were plotted against these LT₅₀ values. Linear regressions were performed using GraphPad PRISM version 10.2.3 (GraphPad Software, Boston, MA, USA, www.graphpad.com) to fit equations to these data points. Statistical differences between the LT₅₀ vs ln(dose) equations of different bacterial strains were determined using PRISM to generate non-overlapping 95% confidence intervals (CIs) of LT₅₀ values at doses of 2,000 or 60,000 CFU.

RESULTS

Testing multiple doses of P. aeruginosa has several advantages in estimating the time to death of G. mellonella following infection with P. aeruginosa

As a first attempt at adopting a time-to-death strategy to measure P. aeruginosa virulence in G. mellonella, we injected a single bacterial dose into larvae and monitored mortality over time. We used OD₆₀₀ measurements to prepare approximate P. aeruginosa inoculum sizes for injection and then performed plating and colony enumeration of the inoculum to more definitively determine these inoculum sizes—a quantification method that did not yield results until the day after the G. mellonella larvae were injected. We encountered two difficulties in using this approach. First, each dose measurement was associated with substantial error, and inoculums of small sizes were prone to especially large relative measurement errors (Fig. S1). These errors had the potential to translate into substantial differences in the times that larvae died, causing problems with reproducibility. Second, optical density vs CFU correlations varied significantly from strain to strain of P. aeruginosa, making it difficult to accurately prepare inoculums of the same targeted sizes to compare the lethality of distinct P. aeruginosa strains (Fig. S2).

To address these issues, we chose to use a range of doses for each tested strain rather than a single targeted dose, thus minimizing the effect of error associated with any one dose. We then plotted the LT₅₀ values vs the natural log of the doses for each strain. An advantage of this approach is that the use of multiple data points allows more accurate estimations of LT₅₀ values for very small doses of bacteria. A second advantage is that a regression line fit to multiple data points allows for the estimation of an LT₅₀ value for any dose and therefore precludes the need to accurately prepare a dose of a targeted size. For example, one may wish to know the LT₅₀ value corresponding to an inoculum of 2,000 CFU. As described above, it can be quite difficult to prepare a bacterial suspension of exactly 2,000 CFU (Fig. S2). However, the regression line fit to the LT₅₀ vs ln(dose) data points generates an equation that provides LT₅₀ values as a function of ln(dose). The LT₅₀ value corresponding to a dose of 2,000 CFU can be estimated from this equation even if the operator was unable to prepare an injection dose of exactly 2,000 CFU.

To illustrate this time-to-death methodology, we applied it to the P. aeruginosa bloodstream isolate PABL089. Ten larvae were infected with a dose (12,300 CFU) of PABL089, and larval death was monitored each hour. An LT₅₀ value (~14 hours) was then estimated based on the time of death of the 10 larvae (Fig. 1A). This process was repeated with additional doses of PABL089 (485, 705, 790, 1,265, 1,650, 2,867, 3,250, 4,350, and 5,950 CFU). The resulting LT₅₀ values were plotted against the natural log of the CFU of each dose (Fig. 1B). The LT₅₀ value of PABL089 at an arbitrarily chosen dose could then be calculated from these data (e.g., 18.2 hours for a dose of 2,000 CFU, as shown in Fig. 1B). This example illustrates how the time-to-death methodology can be used to quantify the virulence of P. aeruginosa strains.

Fig 1.

Estimating the virulence of a P. aeruginosa strain using the time-to-death approach. P. aeruginosa strain PABL089 was used as an example to demonstrate the LT₅₀ approach. (A) Ten G. mellonella larvae were injected with each of the P. aeruginosa inoculums of 12,300, 2,870, or 705 CFU. Cumulative larval mortality was monitored over time. LT₅₀ values were calculated by fitting a sigmoidal curve to mortality observations. (B) LT₅₀ values were next measured for additional doses of PABL089, and all LT₅₀ data points were plotted against the natural log of their respective doses. Colored dots represent the corresponding doses shown in panel A. A linear regression line was fit to the data points. The virulence of PABL089 at any specific dose was then estimated by interpolating the LT₅₀ value corresponding to that dose (18.19 hours at 2,000 CFU in this example).

The weights of G. mellonella larvae affect time to death following inoculation with P. aeruginosa

To further enhance reproducibility, we studied the impact of larval weight on LT₅₀ values. Prior reports have used larvae of various weights for infections, both in the context of P. aeruginosa (20) and other pathogens (21). However, to our knowledge, the impact of weight on virulence readouts has not been systematically investigated. Larvae weighing less than 150 mg were difficult to inject due to their small size and were not tested. Relatively few larvae weighed >300 mg, so those between 300 and 350 mg were combined with those weighing 250–299 mg, and those extremely rare larvae >350 mg were excluded. We injected a series of doses of PABL089 into larvae of the following weight ranges: 150–199, 200–249, and 250–350 mg. LT₅₀ values were plotted against the natural log of dose for each weight range (Fig. 2). The differences between the groups were not statistically significant, although the smaller larvae showed a trend toward dying slightly faster. We chose to exclude larvae weighing less than 200 mg and used larvae weighing between 200 and 350 mg for subsequent experiments.

Fig 2.

Effect of G. mellonella larval weight on LT₅₀ values following injection with P. aeruginosa. Larvae weighing 150–199, 200–249, and 250–350 mg were infected with strain PABL089, and LT₅₀ values were measured. Each symbol represents the LT₅₀ value calculated from 10 larvae given the same dose of PABL089. The experiment was performed twice, and the results were combined. Regression lines were then fit to the data. For the 150–199, 200–249, and 250–350 mg groups, 95% CIs for LT₅₀(2,000 CFU) were 14.73–17.20, 15.96–17.68, and 17.04–17.93 hours, respectively; 95% CIs for LT₅₀(60,000 CFU) were 3.89–15.94, 7.21–15.58, and 10.26–14.57 hours, respectively.

Small variations in the site of injection do not affect time to death of G. mellonella following inoculation with P. aeruginosa

We next examined whether slight variations in the site of injection (e.g., due to variability in the technique used by different operators) altered time-to-death measurements, perhaps by causing traumatic injury to critical G. mellonella structures. In previous studies, G. mellonella were most commonly injected at the final proleg (see the supplemental Methods Protocol, Fig. 4) (21, 22). We examined whether injections with P. aeruginosa strain PABL089 slightly (approximately 1 mm) cephalic or caudal to the final proleg or at the site of the proleg affected LT₅₀ values. The LT₅₀ vs ln(dose) values were similar for the different injection sites, indicating that slight variations in injection sites did not appreciably affect virulence measurements (Fig. 3). For the purpose of consistency, all injections in subsequent experiments were targeted directly to the final proleg.

Fig 3.

Effect of injection site on LT₅₀ values following inoculation with P. aeruginosa. P. aeruginosa strain PABL089 was injected ~1 mm cephalic (“cephalic”), directly at (“Last Proleg”) or ~1 mm caudal (“Caudal”) to the larval final proleg, and LT₅₀ values were measured. Each symbol represents the LT₅₀ value associated with 10 larvae. The experiment was performed twice, and the results were combined. Regression lines were fit to the data and are shown. The differences between the three injection sites were not statistically significant. For the proleg, caudal, and cephalic injection groups, 95% CIs for LT₅₀(2,000 CFU) were 17.69–19.26, 17.67–20.02, and 17.11–20.78 hours, respectively; 95% CIs for LT₅₀(60,000 CFU) were 11.95–15.02, 11.23–15.82, and 8.63–15.78 hours, respectively.

Testing of multiple doses allows for the assessment of the quality of time-to-death measurements

Another advantage of testing multiple doses rather than a single dose for each strain is that the results from multiple doses can be used as a quality control metric. Despite standardization of the injection process, as described above, we noted that unexpectedly short or long time-to-death measurements of G. mellonella larvae sometimes still occurred, especially when inexperienced operators performed the experiments. An example of such a set of data that was generated with the bloodstream isolate PABL051 is shown in Fig. 4A. We took advantage of the theoretical linear relationship between LT₅₀ and ln(dose) to generate a quality control metric that would flag such data sets. For each data set, the scatter in LT₅₀ vs ln(dose) was quantified by calculating an R² coefficient for the goodness-of-fit of an optimal regression line. Those data sets with R² coefficients < 0.60 were deemed low quality, discarded, and subsequently repeated (Fig. 4B). In this way, we minimized the impact of erroneous data and statistical noise on virulence measurements.

Fig 4.

Use of R² as a quality control metric. Different doses of strain PABL051 were injected into larvae, and the resulting LT₅₀ values were calculated. Each symbol represents 10 larvae. (A) In the initial experiment, data points fit poorly to a regression line (R² of 0.125), suggesting substantial error in the measurements. These data were discarded. (B) The experiment was repeated, and subsequent results showed much better alignment to the expected linear relationship (R² of 0.91). The second data set was deemed sufficiently accurate.

The time-to-death approach can be used to compare the virulence of P. aeruginosa strains

We next examined the utility of using the optimized LT₅₀ approach to compare the virulence of different P. aeruginosa strains. In addition to the bloodstream isolates PABL051 and PABL089, which were used in the preceding experiments, we tested four other bloodstream isolates: PABL038, PABL066, PABL083, and PABL093. These six isolates were chosen because they had previously been tested in a mouse model of bacteremia and shown to differ in their virulence (17, 18). The LT₅₀ methodology was applied to each of the six isolates, and LT₅₀ vs ln(dose) curves were generated (Fig. 5, which includes the PABL089 data shown in Fig. 1). Inspection of these curves suggested that lower bacterial doses better discriminated between strain virulence than higher doses. For example, at a dose of 2,000 CFU, the LT₅₀ values of the six strains varied from 12.6 to 18.2 hours, with PABL089 and PABL066 having statistically significantly higher LT₅₀ values and PABL083 having a statistically significantly lower LT₅₀ value than some of the other strains (Table 1 and Fig. 5; statistically significant differences are noted in the legend of Fig. 5). In contrast, at a theoretical dose of 60,000 CFU, the LT₅₀ values of the six strains ranged from 8.4 to 11.8 hours, with only PABL066 having an LT₅₀ value statistically significantly different from some of the other strains (Table 1 and Fig. 5; statistically significant differences are noted in the legend of Fig. 5). These findings indicate that a dose of 2,000 CFU better discriminated the virulence differences of these six strains than a dose of 60,000 CFU. For subsequent virulence comparisons, we used a dose of 2,000 CFU.

Fig 5.

Virulence of six strains of P. aeruginosa in G. mellonella. LT₅₀ vs ln(dose) data were generated for six bloodstream isolates of P. aeruginosa. Regression lines were then fit to each set of data points. Each data point represents the results of a single experiment in which 10 larvae received a single dose of bacteria. Each experiment was performed using at least five different bacterial doses. Each experiment consists of at least two biological replicates performed by different operators. Cumulative results are displayed. At a dose of 2,000 CFU, statistically significant differences (non-overlapping 95% CIs) were observed between PABL089 and PABL051, PABL066, PABL083, PABL093, PABL038; between PABL066 and PABL083, PABL093; between PABL083 and PABL051; between PABL093 and PABL051. At a dose of 60,000 CFU, statistically significant differences were observed between PABL066 and PABL093, PABL083, PABL038.

TABLE 1.

Virulence measurements of P. aeruginosa bloodstream isolates in G. mellonella and mice

Strain name	G. mellonella LT50 at 2,000 CFU (hours)	G. mellonella LT50 at 60,000 CFU (hours)	Murine LD50 (CFU)a
PABL083	12.6	10.6	6.4
PABL093	13.2	9.9	6.3
PABL038	14.3	8.4	7.4
PABL051	14.7	10.9	7.5
PABL066	15.9	11.8	7.7
PABL089	18.2	9.5	7.3

^a Mouse LD₅₀ values were previously reported (17, 18).

P. aeruginosa strain virulence rankings in G. mellonella show modest agreement with those obtained using a mouse model of bacteremia

We compared the LT₅₀ virulence rankings of the six strains using G. mellonella to those previously reported using a mouse bacteremia model (Table 1). There was a modest correlation between mouse log(LD₅₀) values and G. mellonella LT₅₀(2,000 CFU): R² of 0.46 (Fig. 6). This finding confirms other reports suggesting that G. mellonella is a moderately useful surrogate for mice in predicting the virulence potential of P. aeruginosa (12). However, to our knowledge, this is the first investigation into the correlation between the G. mellonella model and the murine model for genetically diverse clinical P. aeruginosa isolates.

Fig 6.

Comparison of the virulence of six P. aeruginosa strains in the G. mellonella model and the murine model. Virulence measures in the G. mellonella model (LT₅₀ values at a dose of 2,000 CFU) and a murine model of bacteremia (LD₅₀ values) were plotted, and a moderate level of agreement was observed (R² = 0.46).

In G. mellonella, the time-to-death approach has advantages over the LD₅₀ approach in measuring the virulence of P. aeruginosa

We have mentioned several advantages of the LT₅₀ approach over the conventional LD₅₀ approach. To illustrate this, we applied the LD₅₀ approach to measure the virulence of PABL038, a P. aeruginosa strain with a moderate level of virulence in both G. mellonella and mice (Table 1). G. mellonella larvae were infected with P. aeruginosa doses ranging from 22 to 580 CFU, and mortality was measured at 18- and 24-hours post-inoculation. Even when bacterial numbers as low as ~22 CFU were injected, all or nearly all the larvae died, making it difficult to estimate LD₅₀ values (Table 2). In contrast, the LT₅₀ approach successfully captured the virulence of this strain (Table 1). This example illustrates the utility of the LT₅₀ approach in measuring the virulence of P. aeruginosa.

TABLE 2.

Number of dead larvae following injection of different-sized inoculums of strain PABL038^a

	Inoculum size
Hours post-infection	580 CFU	265 CFU	160 CFU	40 CFU	22 CFU
18	10	10	9	9	10
24	10	10	9	9	10

^a Total of 10 larvae were injected for each group.

DISCUSSION

G. mellonella is frequently used as a model organism for measuring bacterial virulence and response to antibiotics, but the highly pathogenic nature of P. aeruginosa toward G. mellonella poses special challenges. We provide an optimized protocol for the use of time-to-death as a virulence measure for P. aeruginosa infections of these larvae. We evaluated the impact of variables such as weight and injection site on this virulence readout. We also optimized an analytical approach that uses these results to allow for strain-to-strain comparisons even when relatively large uncertainties exist in dose estimations. Finally, we provide a quality control metric that is useful in ensuring that only reliable data are included in the final analysis. This protocol was tested on six P. aeruginosa clinical isolates and yielded results with moderate agreement to those previously obtained using a mouse model. This protocol may be useful to the P. aeruginosa research community.

Much of the previous work testing P. aeruginosa in G. mellonella has used LD₅₀ as a measure of virulence (11, 12, 15, 23–25). Many of these reports compared parental strains of P. aeruginosa to mutant strains that were defective in the production of critical pathogenic factors and were, therefore, highly attenuated in virulence. The LD₅₀ approach is well suited for such comparisons. However, we found it performed suboptimally when comparing strains that had only small differences in virulence. For example, the strains we investigated killed nearly all larvae over a period of 18–24 hours even with bacterial doses as low as 22 CFU (Table 2), making it difficult to determine an LD₅₀ value. In contrast, the LT₅₀ approach distinguished virulence differences between some of these strains. A second approach favored by investigators testing the efficacy of antimicrobial compounds is to compare Kaplan-Meier survival curves of infected G. mellonella larvae exposed to antibiotics or vehicles (26–28). This focus on survival differences between treatment and non-treatment cohorts circumvents the issues associated with measuring LD₅₀ values. However, this approach (which in essence is an LT₅₀ measurement for a single dose) works less well when comparing the survival of multiple P. aeruginosa strains because it requires that the same inoculum of each strain be given to the cohorts of larvae. It becomes difficult to determine whether measured differences in virulence between two strains are biological or are instead due to small differences in the inoculum sizes or to the experimental uncertainties in the measurements. The LT₅₀ approach detailed here accurately captures small differences in virulence between even highly pathogenic strains by using multiple measurements and allowing for interpolation of time-to-death for any inoculum size.

A comparison of the virulence of different strains unexpectedly uncovered an interesting aspect of P. aeruginosa virulence. We tested six P. aeruginosa bloodstream isolates previously shown to have a wide range of virulence in a mouse model of bacteremia (17, 18). The LT₅₀ values of these six strains were more distinct at low inoculums (~2,000 CFU) than at high inoculums (~60,000 CFU). In other words, at a high inoculum, all six strains killed G. mellonella at about the same time, whereas at a low inoculum, they differed somewhat in the time required to cause death. What strain differences, such as quorum sensing, biofilm formation, or growth rate, account for these dose-dependent effects is an interesting avenue of future study and may inform P. aeruginosa pathogenesis.

A critical measure of the usefulness of the G. mellonella model for the study of P. aeruginosa is how closely virulence measurements in G. mellonella reflect those obtained in mice. Although G. mellonella lacks an adaptive immune response, this organism’s innate immune system has similarities with those of vertebrates (2). Therefore, it has been proposed that G. mellonella may be quite useful for modeling acute P. aeruginosa infections, in which the innate immune response is critical (29). Results from the G. mellonella model were moderately similar to those from mice (Table 1). Although these observations are intriguing, studies using many more P. aeruginosa strains obtained from a variety of clinical and environmental sources will be necessary to demonstrate the true usefulness and limitations of the G. mellonella model. Similarly, the degree of agreement between G. mellonella and mice may also depend on the type (breed, immunocompromised status, and site of infection) of the mouse model used.

Our study has several limitations. As mentioned above, only bloodstream isolates of P. aeruginosa were used, and other types of isolates (pneumonia, urinary tract infections, and natural environment) will need to be tested to show the generalizability of the approach. In particular, the use of cystic fibrosis isolates with this LT₅₀ methodology will be important, as these isolates differ markedly in their virulence compared to acute infection isolates (25). We applied the protocol to only six strains. A much larger collection of strains will need to be tested to more definitively determine how well this G. mellonella model ranks P. aeruginosa strain virulence compared to mice. Finally, we purchased G. mellonella larvae from a non-scientific commercial vendor, and the genetics and breeding history of these larvae are unclear. We cannot be certain that other groups using G. mellonella from different sources will obtain results similar to ours. The purchase of inbred G. mellonella, which are currently available in Europe (30) but not (to our knowledge) in the United States, would circumvent this issue.

In summary, we provide a protocol for reproducibly measuring and comparing the virulence of P. aeruginosa using G. mellonella. This approach may be useful in increasing our understanding of P. aeruginosa pathogenesis and the mechanisms by which this bacterium causes disease. It may also be applicable to other bacterial species, especially those that are highly virulent in G. mellonella (e.g., Burkholderia pseudomallei and Burkholderia mallei [31]) and for which the lethal dose methodology is not optimal. In addition, this methodology should be applicable to efficacy studies for testing antibiotics and alternative therapeutics.

SUPPLEMENTAL MATERIAL

The following material is available online at https://doi.org/10.1128/spectrum.01666-24.

Supplemental figures. spectrum.01666-24-s0001.pdf.

Fig. S1 and S2.

Supplemental methods. spectrum.01666-24-s0002.pdf.

Protocol for performing Galleria experiments.

OPEN PEER REVIEW. reviewer-comments.pdf.

An accounting of the reviewer comments and feedback.

ASM does not own the copyrights to Supplemental Material that may be linked to, or accessed through, an article. The authors have granted ASM a non-exclusive, world-wide license to publish the Supplemental Material files. Please contact the corresponding author directly for reuse.