Emerging Pathogens in Urinary Tract Infections: Virulence and Phenotypic Characterization of Pseudomonas aeruginosa strains
University of South Alabama, College of Medicine, Department of Microbiology and Immunology, 5851 USA Drive N, Mobile, Alabama 36688, USA
Department of Biological Sciences, College of Medicine and Health Sciences, Khalifa University, Abu Dhabi, UAE
Center for Biotechnology (BTC), Khalifa University, Abu Dhabi, UAE
Moderna TX, Department of Infectious Disease Research, 325 Binney, Cambridge, Massachusetts 02142, USA
Bruker Spatial Biology, 3350 Monte Villa Parkway, Bothel, Wisconsin 98021, USA
#Address correspondence to A.E. Shea, aeshea@southalabama.eduABSTRACT
Urinary tract infections (UTIs) affect a broad patient population and inflict a substantial financial burden on the U.S. healthcare system. While uropathogenic Escherichia coli (UPEC) causes the majority of cases, other pathogens are emerging. Analysis of patient data from our healthcare system in the Gulf Coast region of Alabama revealed that Pseudomonas aeruginosa accounted for 4.0% of UTI cases, roughly double the national average, prompting further investigation into this historically understudied uropathogen. Here, we performed whole-genome sequencing and phenotypic assays on 55 urinary P. aeruginosa isolates to identify key drivers of pathogenicity in the context of UTI. Multilocus sequence typing identified 19 novel sequence types, underscoring the uncharacterized diversity of urinary P. aeruginosa isolates. Serotype O6 was most common and enriched in patients with indwelling catheters, whereas O4 was linked to diabetes mellitus. Antibiotic susceptibility testing (AST) revealed high levofloxacin resistance (30.9%), with 23.6% multidrug-resistant (MDR) and 9.1% extensively drug-resistant (XDR) isolates. Resistance patterns correlated with demographics, including significantly higher meropenem and aztreonam resistance in isolates from African American patients. Phenotypic assays of growth, motility, and biofilm formation revealed negative correlations between antibiotic resistance and virulence. Specific virulence genes predicted enhanced iron acquisition, hemolysis, and colonization potential. Notably, motility and exotoxin profiles emerged as strong predictors of P. aeruginosa ascension in a murine UTI model. Together, these findings provide new biological and clinical insight into P. aeruginosa as a uropathogen and emphasize the need for continued research.
IMPORTANCE
Pseudomonas aeruginosa is an emerging but understudied pathogen in urinary tract infections (UTIs). Given its resilience, adaptability, and the growing threat of multidrug resistance, P. aeruginosa remains a significant challenge in clinical microbiology and infection control. Our data reveal an increased prevalence of P. aeruginosa in our local patient population. In this study, we examined both genotypic and phenotypic traits of clinical isolates and correlated them with colonization in murine models and extensive patient metadata. We identified strong associations between antibiotic resistance patterns and patient demographics. Novel sequence type strains were linked to motility phenotypes in vitro. Additionally, specific flagellar alleles were associated with enhanced murine kidney colonization and recurrent UTIs in patients. These findings provide new insight into the evolutionary adaptations that contribute to P. aeruginosa uropathogenicity and support a more nuanced understanding of its clinical significance.
INTRODUCTION
Pseudomonas aeruginosa is a Gram-negative, opportunistic pathogen best known for causing severe lung and wound infections, yet it can also pose significant challenges in the urinary tract1. Although urinary tract infections (UTIs) are predominantly caused by uropathogenic Escherichia coli (UPEC), P. aeruginosa accounts for up to 10% of catheter-associated urinary tract infections (CAUTIs) and 16% of UTIs in intensive care units (ICUs)2–4. The prevalence of P. aeruginosa in hospital-acquired infections is driven by innate resistance to many antibiotics, adaptability to diverse environments, and numerous virulence factors. However, in contrast to extensive studies in lung and wound infection models, the characteristics that allow P. aeruginosa to successfully colonize the urinary tract remain poorly understood.
A defining feature of P. aeruginosa is its extensive intrinsic and acquired antibiotic resistance mechanisms. Its large genome encodes multiple efflux pumps, β-lactamases, and aminoglycoside-modifying enzymes, conferring innate resistance to a broad range of antimicrobial classes5. Pseudomonas clinical isolates are often multi-drug resistant (MDR), defined as resistance to ≥1 agent in ≥3 antimicrobial classes, or extensively drug-resistant (XDR), defined as resistance to all but one or two classes6. Approximately 29% of complicated UTIs (cUTIs) caused by P. aeruginosa are due to MDR strains, making treatment especially challenging7. Because of this, the Centers for Disease Control and Prevention (CDC) has classified MDR P. aeruginosa as a serious public health threat8.
In addition to antimicrobial resistance, P. aeruginosa employs a broad range of virulence mechanisms. Biofilm formation, particularly on urinary catheters, protects bacteria from host defenses as well as antibiotic penetration and relies on flagellar motility genes (type A flaA and type B fliC) and pili genes such as pilA1,9. Quorum sensing (QS) further coordinates biofilm formation and persistence by linking environmental sensing to virulence expression via key transcriptional regulators such as lasR and rhlR10. Given the rise of antibiotic resistance, these virulence mechanisms have emerged as targets of antivirulence therapies11,12. Indeed, experimental disruption of QS pathways has been shown to significantly reduce virulence in the urinary tract13,14 , highlighting the importance of researching these mechanisms to better understand and treat P. aeruginosa UTIs.
Another critical virulence feature is the secretion of exotoxins, particularly via the type III secretion system (T3SS). The T3SS injects cytotoxins such as ExoS, ExoT, ExoU, and ExoY directly into host cells, disrupting cytoskeletal integrity, interfering with immune signaling, and promoting cell death15–18. Interestingly, the T3SS was shown to be essential for establishing acute infection but not required for persistence in chronic infection in a murine model of CAUTI19. Among these four exotoxins, ExoU and ExoS are considered most clinically relevant and are thought to be mutually exclusive in P. aeruginosa strains20,21. ExoU is particularly associated with virulence, severe cytotoxicity, and poor clinical outcomes in the context of pneumonia22,23. In addition to the T3SS, the type II secretion system (T2SS) also contributes to P. aeruginosa virulence via secretion of exotoxin A (ToxA), a highly cytotoxic protein strongly linked to pneumonia and chronic lung infections in cystic fibrosis patients24–26. Expression of exotoxin A is directly regulated by the iron-scavenging siderophore pyroverdine, partially encoded by pvdA, which is essential for growth in iron-limited environments such as urine27–29. No single virulence factor alone drives P. aeruginosa pathogenesis, but rather it is the combination of these factors that enables successful infection of the urinary tract.
In addition to virulence factors, variability in P. aeruginosa serotypes can also influence infection severity and treatment outcomes. Among the 20 known P. aeruginosa O-antigen serotypes, O6 and O11 have been associated with increased mortality in patients with pneumonia30. Isolates of serotype O11 often harbor exoU and cause significant epithelial damage and have also been associated with multidrug resistance (MDR)31,32. Given their pathogenic potential, recent studies have investigated O-antigens as vaccine therapeutic targets, underscoring the importance of characterizing O-antigen serotypes in clinical P. aeruginosa isolates33. Patient-related factors also play a critical role in UTI susceptibility and recurrence. Several factors are known to increase the risk of P. aeruginosa UTIs, including indwelling urinary catheters (IDCs), immunosuppressive therapies, diabetes mellitus (DM), and prolonged hospitalization7. However, the roles of other host factors in P. aeruginosa infection and virulence in the urinary tract remain relatively understudied, highlighting the need for integrative studies that investigate both pathogen diversity and patient characteristics in the context of UTI.
Here, we phenotypically and genotypically assessed 55 P. aeruginosa clinical urinary isolates from a diverse patient population. We performed bacterial growth, biofilm formation, iron acquisition, cytotoxicity, and motility assays to directly assess P. aeruginosa virulence and persistence in vitro. We also conducted whole-genome sequencing (WGS) to determine the presence of important virulence and antibiotic resistance genes as well as characterize the multi-locus sequence type (MLST) and O-antigen serotype of each strain. With these results and abundant patient metadata, we correlated phenotype, genotype, and patient variables to uncover drivers of P. aeruginosa pathogenicity in UTI, validated by our in vivo murine UTI model. Understanding the interplay between these factors is crucial for developing effective therapeutic and UTI preventive strategies.
RESULTS
Emergence of novel genotypes in uropathogenic P. aeruginosa strains
Pseudomonas aeruginosa has been extensively studied in the context of lung and wound infections but relatively understudied as a urinary tract infection (UTI) pathogen. Nationwide, P. aeruginosa accounts for only 1% of uncomplicated UTIs and 2% of complicated UTIs3; however, we report a prevalence of 4.03% in our local healthcare system (Fig. S1). This discrepancy prompted us to investigate the genotypic and phenotypic characteristics of P. aeruginosa strains in our region and to explore their associations with patient variables to better understand clinical risk. Fifty-five P. aeruginosa clinical isolates were obtained from urine samples of patients within the University of South Alabama healthcare system, representing a diverse cohort with various comorbidities and known UTI risk factors (Table 1). Three reference strains were included in this study for comparison: PAO1, PA103, and PA27853, isolated from wound, lung, and bloodstream infections, respectively34–37. Indeed, there is no available P. aeruginosa urinary type strain for comparison.
Whole-genome sequencing (WGS) was performed on all clinical isolates, and the genomes were assembled and annotated using BV-BRC38. A phylogenetic tree constructed from single nucleotide polymorphisms (SNPs) revealed multiple distinct clades, suggesting a high degree of genomic diversity among clinical isolates (Fig. 1A). Multilocus sequence typing (MLST) via PubMLST39 identified 19 novel sequence types (STs) (Table 2), further demonstrating the lack of genotypic characterization of uropathogenic P. aeruginosa. O-antigen serotyping was performed with the PAst in silico online tool40, and a total of 8 distinct serotypes were identified in our cohort, with O6 being the most prevalent (38.18%) (Fig. 1B). The O6 serotype was enriched in patients with indwelling catheters (IDCs) (Fig. 1C), while O4 strains were more common in patients with diabetes mellitus (DM) (Fig. 1D). These findings demonstrate previously uncharacterized diversity in urinary isolates, motivating further investigation into how this genetic variation shapes antibiotic resistance and virulence phenotypes.
Patient demographics strongly associate with specific antibiotic resistance profiles in P. aeruginosa UTI strains
Antibiotic resistance is becoming an urgent global threat, and P. aeruginosa is a noted pathogen of high importance (ESKAPE pathogen) due to its innate and acquired resistance to many antibiotics41. First-line therapies often include the cephalosporins cefepime and ceftazidime, or the fluoroquinolones ciprofloxacin and levofloxacin, while meropenem is often reserved for highly resistant strains42. In our cohort, antimicrobial susceptibility testing (AST) revealed a wide range of resistance profiles to seven clinically relevant antibiotics (Fig. 2A). Levofloxacin resistance was the most common, with 36.4% (20/55) of isolates conferring resistance. Of the 55 strains, 13 (23.6%) were multidrug-resistant (MDR) and 5 (9.1%) were extensively drug-resistant (XDR). Isolates of the O11 serotype were 7.9 times more likely to be XDR (Fig. S2). Interestingly, resistance patterns varied by specific patient demographics and comorbidities. Isolates from African American patients were significantly more likely to be resistant to meropenem (Fig. 2B) and 14.2 times more likely to have intermediate aztreonam resistance (Fig. 2C) compared to those from White patients. Intermediate cefepime resistance was more common among strains isolated from polymicrobial infections than single causative agent cases (Fig. 2D), and aztreonam resistance positively correlated with the presence of novel sequence types (Fig. 2E). In addition, isolates from male patients harbored significantly more resistances per strain than those from female patients (Fig. 2F), with 50% of male-derived isolates and 24% of female-derived isolates having 1 or more resistance. To complement phenotypic testing, we also assessed antibiotic resistance genotypically using the Comprehensive Antibiotic Resistance Database (CARD)43 (Fig. S3) and discovered the presence of the MDR-associated gene armR was negatively correlated with patient BMI (Fig. 2G, Fig. S4). Together, these patterns point to complex drivers of resistance that span bacterial genotype, host demographics, and infection context.
Virulence-associated genes predict differential growth in human urine
To assess bacterial fitness under both nutrient-rich and host-relevant conditions, we quantified the growth of 55 P. aeruginosa clinical isolates and 3 reference strains in Luria Broth (LB) and filter-sterilized, pooled human urine. Many clinical isolates outperformed the reference strains in either medium, but none excelled in both LB and human urine, suggesting a metabolic trade-off (Fig. 3A). The presence of virulence genes was then determined by Basic Local Alignment Search Tool Protein (BLASTP), and these results were correlated with growth. Because gene detection was based on similarity to PAO1 alleles, some genes classified as “absent” may instead represent divergent urinary variants with <85% amino acid identity. Growth in human urine was enhanced 1.2-fold in isolates harboring AMR-associated genes parS and armR and 1.5-fold in isolates with the virulence gene toxA (Fig. S5A). Additionally, strains with the O5 serotype showed a 20.7% growth increase (Fig. S5B) while strains with serotype O11 exhibited a 28.9% growth decrease in human urine (Fig. S6C) compared to other serotypes. Strains isolated from polymicrobial UTIs showed a 36.6% increase in human urine growth (Fig. 3B) compared to isolates from single organism infections. In LB, strains with virulence factors exoU (Fig. 3C) and aprA (Fig. 3D) as well as those with pili-associated gene pilA (Fig. 3E) displayed significantly reduced growth. Interestingly, no patient-associated variables aside from co-infection correlated with pathogen growth (Fig. S6). Overall, these findings indicate that specific virulence factors not only affect pathogenic potential but may also shape metabolic adaptability in the urinary tract.
Exotoxin genes predict hemolytic activity in clinical isolates
Iron acquisition and cytotoxicity are critical determinants of P. aeruginosa pathogenicity, influencing both survival in the iron-limited urinary tract and the extent of host tissue damage1,29. To assess these traits among our clinical isolates, we quantified iron acquisition using chrome azurol S (CAS) plates and hemolytic activity using blood agar. Most strains demonstrated either strong iron acquisition or hemolytic activity, but strain SL819 was an exception that displayed increased activity in both assays (Fig. 4A). Siderophore production was significantly increased in strains carrying virulence factors aprA (Fig. 4B) and pvdA (Fig. 4C), by 16.1% and 6.1%, respectively. In contrast, isolates with novel sequence types (Fig. 4D) showed a 6.2% decrease in iron chelation compared to known MLST strains. Hemolytic activity was enhanced by 18.4% in isolates with exoU (Fig. 4E) and 44.5% in those encoding rhlR (Fig. 4F), consistent with the known roles of these genes in blood cell lysis44,45. In contrast, strains resistant to levofloxacin, aztreonam, or meropenem demonstrated reduced hemolytic activity (Fig. 4G), suggesting a potential trade-off between cytotoxicity and antibiotic resistance. Furthermore, O6 serotype strains were associated with 6.6-fold higher blood urine concentrations than other serotypes (Fig. S7). Collectively, these findings highlight that serotype and toxin gene presence predict host damage and may also influence the urinary niche, while resistance traits may reduce cytotoxic ability.
Motility is enhanced among novel P. aeruginosa UTI isolates.
Motility and biofilm formation are central to P. aeruginosa persistence in the urinary tract, allowing tissue colonization and evasion of antibiotic clearance1,9,46. To evaluate these traits, we assessed swimming motility and biofilm production in vitro across our clinical isolates. Despite all isolates encoding the flagellar machinery, 5 isolates were non-motile, while the rest demonstrated varying levels of motility (Fig. 5A). Isolates from patients with recurrent urinary tract infections (rUTI) were 5.5 times more likely to carry the type A flagellin gene (flaA), whereas isolates from non-rUTI patients were more likely to carry type B flagellin (fliC) (Fig. 5B). Importantly, isolates belonging to novel sequence types exhibited 20.2% greater swim motility compared to those with established sequence types (Fig. 5C). Increased motility was associated with the presence of the quorum-sensing regulator rhlR (Fig. 5D), which also affects biofilm formation. Biofilm formation was 2.2 and 2.4 times higher in strains of the O3 serotype (Fig. 5E) and strains harboring the transcriptional regulator lasR (Fig. 5F), respectively. In contrast, isolates resistant to levofloxacin exhibited 51.8% weaker biofilm formation than those that were susceptible (Fig. 5G). These trends suggest that motility and biofilm capacity may act as complementary strategies for persistence in the urinary tract.
Swimming motility and ExoS drive colonization and persistence of uropathogenic P. aeruginosa in the murine model of UTI
To complement our in vitro assay findings, we used the traditional ascending murine UTI model to directly assess how phenotypic and genotypic characteristics contribute to colonization and persistence in vivo47. Four P. aeruginosa strains (PAO1, PA103, SL158, and SL192) were used to infect CBA/J mice (n=20) via transurethral inoculation.
Strains were selected to represent distinct Type III secretion system toxin profiles, enabling comparison between exoS-dominant and exoU-dominant genotypes. Urinary bacterial burden was quantified over a 96-hour period through serial CFU enumeration from urine samples, followed by organ collection and plating at the endpoint to determine bacterial load in the bladder, kidneys, and spleen. PAO1 consistently maintained the highest urine CFU burden, whereas PA103 exhibited a decline in CFU/mL over time (Fig. 6A, Fig. S8). A similar pattern was seen in tissue CFU burdens, where PA103 was unable to establish organ colonization, the two clinical isolates displayed intermediate levels, and PAO1 colonized very well (Fig. 6B). Kidney colonization differed by motility phenotype (Fig. 6C), with motile strains colonizing the kidney at median levels over one log higher than non-motile strains. Quantification of organ and urine CFU revealed differences in colonization capacity associated with the presence of key virulence genes (Fig. 6D). Strains lacking the toxin gene exoS exhibited a 100-fold decrease in the urine and a 5-fold decrease in both bladder and kidney colonization compared to exoS-positive strains (Fig. 6D). Additionally, isolates expressing type B flagellin (fliC) showed a 10.9-fold advantage in bladder colonization and a 13.5-fold advantage in kidney colonization relative to those expressing type A flagellin (flaA). To contextualize these in vivo findings, we compared the prevalence of these same virulence genes across P. aeruginosa genomes from other infection sources from the PATRIC database (Fig. 6E). We found that UTI isolates were significantly enriched for exoS compared to lung isolates. Collectively, these findings indicate motility and exotoxin ExoS are key fitness factors essential to P. aeruginosa colonization, persistence, and ascension in the urinary tract.
DISCUSSION
In this study, we performed a comprehensive genotypic and phenotypic characterization of Pseudomonas aeruginosa strains isolated from urinary tract infections (UTIs) within our regional healthcare system. While P. aeruginosa has been extensively studied in the context of respiratory and wound infections, its role in UTIs has been understudied by comparison. The absence of a dedicated UTI reference strain underscores this gap. Indeed, our experimental work relied on type strains derived from wound (PAO1), lung (PA103), and bloodstream (PA27853) infections34–37, demonstrating the need for dedicated UTI reference strains and models to advance understanding of P. aeruginosa uropathogenesis.
Our university healthcare system has a higher-than-expected prevalence of uropathogenic P. aeruginosa. Additionally, 34.5% of our isolates collected represented novel sequence types, indicating previously uncharacterized pathogen lineages circulating in our region. These novel isolates displayed a unique phenotypic profile including enhanced motility and reduced iron-chelating activity. Despite these shared phenotypes, phylogenetic analysis revealed high genomic diversity, with no distinct clade encompassing these novel sequence types. Resistance to the monobactam antibiotic aztreonam was also enriched among novel sequence types, proposing that emerging genotypes may be co-acquiring resistance alongside other traits relevant to urinary tract colonization. This degree of genomic novelty suggests the emergence of previously undescribed urinary-adapted lineages, potentially driven by selective pressures unique to our geographic region. Given the known heterogeneity of many P. aeruginosa virulence and resistance loci, these findings further support the need for a UTI-specific prototype strain that reflects the genomic features of urinary lineages.
O-antigen serotyping identified eight distinct serotypes in our cohort, with O6, O11, and O5 comprising the majority of isolates. These serotypes have also been among the most prevalent in ventilator-associated pneumonia and burn wound isolates30,31,48. Interestingly, we found previously undescribed correlations between serotypes and patient data in our cohort. O6 strains were more common in patients with indwelling urinary catheters (IDCs) and were also linked to elevated blood in the urine, while O4 strains were more common in patients with diabetes mellitus (DM). Together, these findings highlight the value of integrating genotypic characterization with patient data to better define patterns of P. aeruginosa infection across diverse patient cohorts.
Although antibiotic resistance rates in our patient population were consistent with those in healthcare-associated infections49, we discovered novel ties between resistance phenotypes and host demographics. Notably, isolates from African American patients exhibited significantly higher resistance to meropenem and aztreonam than those from White patients, suggesting possible population-level differences in prior exposures or environmental reservoirs. Strains of serotype O11 were more likely to be extensively drug-resistant (XDR), consistent with previous studies correlating O11 to antibiotic resistance50. As described in other infection models51,52, we observed a trade-off between biofilm formation and antibiotic resistance. Interestingly, resistance to levofloxacin, aztreonam, or meropenem was associated with reduced hemolytic activity, raising the possibility of previously undescribed fitness trade-offs between resistance and cytotoxicity in the context of uropathogenic P. aeruginosa.
To translate these in vitro phenotype findings in vivo, we utilized the murine model of ascending UTI. To date, there is a paucity of literature utilizing this model with P. aeruginosa, with most prior work focused on catheter-associated infection53,54. Studies in LACA mice have shown quorum sensing systems including PQS, Rhl, and Las are crucial for P. aeruginosa colonization and persistence in the urinary tract13,55. In our CBA/J model, previously only used in the context of P. aeruginosa co-infections56,57, swimming motility was a key predictor of urinary tract colonization. This pattern aligns with observations in other uropathogenic species3,58 and is consistent with the established role of motility in P. aeruginosa dissemination across multiple infection sites, including lung46,59, wound60, and burn61 models.
Flagellin type also influenced colonization outcomes. PAO1, which carries type B flagellin (fliC), exhibited a modest colonization advantage over strains encoding type A flagellin (flaA). Interestingly, flagellin type also correlated with recurrent UTI; flaA was enriched in isolates from patients with recurrent UTIs, whereas fliC was more common in non-recurrent cases. This may indicate selection to avoid immune recognition of FliC by TLR562. Indeed, FliC is very immunogenic and has proven effective as a therapeutic agent against P. aeruginosa in experimental models63–66.
In addition to motility, we also found exotoxin profiles play a role in P. aeruginosa colonization. In a previous study, the type 3 secretion system (T3SS), specifically the injection of ExoU, was found to be the main contributor to virulence in acute catheter-associated urinary tract infection (CAUTI)67. The T3SS is also important in lung and wound pathogenesis, but there is debate over whether exoS or exoU plays a larger role68–70. In our murine UTI model, exoS-positive strains (SL158, PAO1) exhibited significantly higher bacterial burdens than exoU-positive strains (SL192, PA103), suggesting ExoS plays a more prominent role in urinary tract colonization. Supporting this, when analyzing urinary isolates from the PATRIC database, we found that exoS was enriched in urinary P. aeruginosa compared to lung isolates. Larger in vivo cohorts will be necessary for refining these conclusions and clarifying the mechanisms of P. aeruginosa dissemination and persistence in the urinary tract, as has been done in similar studies with E. coli71.
While many previous studies have evaluated PAO1 in the murine UTI model, to our knowledge, none have assessed PA103. Notably, PA103 was the only strain unable to colonize the murine urinary tract, posing the question of what phenotypic and genotypic factors are contributing to this defect. PA103 is non-motile due to a single amino acid substitution in fleQ72, a mutation absent from the other three strains tested in vivo. In fact, none of our clinical UTI strains exhibited this mutation, further demonstrating motility is essential to P. aeruginosa urovirulence. PA103 also possesses a lasR loss-of-function (LOF) mutation that impairs quorum sensing and protease expression, a hallmark commonly seen in chronic P. aeruginosa infections such as cystic fibrosis73–75. In our UTI isolates, lasR-deficient strains demonstrated reduced biofilm formation, consistent with the role of lasR in regulating quorum sensing-associated virulence pathways10. The Las system has indeed been proven crucial to P. aeruginosa pathogenesis in the urinary tract14. Future murine studies restoring motility in PA103 via fleQ repair will provide critical insight into the relative contributions of lost motility versus impaired quorum sensing to its colonization defect.
Overall, our data reveal that P. aeruginosa UTIs in our region are caused by a genetically diverse set of isolates, many of which represent novel lineages with unique combinations of virulence and resistance traits. By integrating genomic, phenotypic, and in vivo analyses, we provide new insights into how P. aeruginosa adapts to and persists within the urinary tract. These findings highlight the need to better recognize P. aeruginosa as a clinically significant uropathogen and to invest in research that addresses this underexplored but clinically important role.
MATERIALS AND METHODS
Bacterial Strains, Media, and Culture Conditions
Bacterial strains used in this study included clinical isolates of Pseudomonas aeruginosa collected under University of South Alabama (USA) IRB protocol #2178590. All clinical isolates were obtained from the University Hospital from the urine culture. Strains were confirmed to be Pseudomonas aeruginosa via an oxidase test and/or Matrix-Assisted Laser Desorption/Ionization Time-Of-Flight (MALDI-TOF). Isolates were coded and stored at -80°C in Luria Broth (LB) containing 20% glycerol. Strains were cultured from a single colony in LB, which contains 0.5 g NaCl, 5 g yeast extract, and 10 g tryptone per liter. To prepare LB agar plates, 5 g of agar was added. Cultures were incubated at 37°C with aeration at 200 rpm unless otherwise stated. In experiments requiring urine, filter-sterilized (0.22 µm) pooled human urine from at least six healthy de-identified female donors was used (deemed IRB exempt by USA).
Biofilm Assay
Biofilm formation was assessed using a crystal violet staining assay. Overnight cultures of strains were diluted 1:100 into LB or filter-sterilized human urine in duplicate in 12-well plates. The plate was sealed with a gas-permeable membrane and incubated statically at 37°C for 24 hours. Wells were washed with 1X phosphate-buffered saline (PBS) and stained with 0.1% crystal violet for 10 minutes. The stain was aspirated, and the wells were washed a second time. Biofilm biomass was quantified by solubilizing the dye with ethanol and measuring absorbance via optical density at 630 nm (OD630) using an Agilent BioTek 800 TS Absorbance Reader.
Iron Acquisition, Hemolysis, and Motility Assays
Iron chelation and siderophore production were measured using the Chrome Azurol S (CAS) assay. CAS agar was prepared as previously described76. For each strain, 5 µL of overnight culture was spotted onto a CAS agar plate and incubated for 16 hours at 37°C. The halo diameter was then measured and recorded in millimeters (mm). Hemolytic activity was evaluated by spotting 5 µL overnight bacterial cultures onto 5% Sheep Blood in Tryptic Soy Agar plates (Hardy Diagnostics CAT# A10). Plates were incubated at 37°C for 16 hours, then zones of clearance were measured in mm.
To assess motility, strains were tested in semi-soft agar. Overnight cultures of each strain were normalized to an OD600 of 10.0, then resuspended in HEPES buffer (pH 8.4). With an inoculating needle, cultures were stabbed into tryptone agar plates with the following composition per liter: 10 g tryptone, 5 g sodium chloride, and 2.5 g agar. Plates were incubated for 16 hours at 30°C, then the swimming diameter was measured in mm.
Growth Curve Assay
Overnight cultures were washed once in 1X PBS, then diluted 1:100 into LB or filter-sterilized, pooled human urine in a 96-well plate. The plate was sealed with a gas-permeable membrane, and bacterial growth was monitored by measuring OD₆₀₀ over time using a BioTek LogPhase600 Microbiology Reader. Readings were taken every 10 minutes for 24 hours.
Murine Model of UTI
Female CBA/J mice between the ages of 6 to 8 weeks old were acquired from Jackson Laboratories. Under ketamine/xylazine anesthesia, mice were transurethrally inoculated with 50 µL of 2 x 108 CFU/mL bacterial suspension of strain PAO1, PA103, SL158, or SL192 with a sterile polyethylene catheter connected to an infusion pump47,77. Urine samples were collected and plated on LB agar every 24 hours post-inoculation to assess bacterial load. After 96 hours, the mice were euthanized, and the bladders, kidneys, and spleens were aseptically removed and homogenized. Homogenates were then serially diluted and plated on LB agar to assess bacterial burden. All protocols were approved by the Institutional Animal Care and Use Committee (IACUC #2006187) at the University of South Alabama.
Whole-Genome Sequencing
Genomic DNA was extracted using the Promega Wizard® Genomic DNA Purification Kit following the manufacturer’s instructions. Libraries were prepared using standard Illumina protocols to produce paired end 150 bp reads. Isolates SL12-SL736 were sequenced by Moderna TX via the NovaSeq 6000 system, and isolates SL754-SL1090 were sequenced by SeqCoast Genomics using the Illumina NextSeq 2000 platform. Raw sequencing data has been deposited in SRA (PRJNA1390275).
Bioinformatics Analysis
Sequencing reads were processed using FastQC78, and genome assemblies were generated with Unicycler79 on BVBRC using an annotated recipe for the P. aeruginosa taxonomy. Eight genomes were flagged as poor quality after this process (SL53, SL157, SL158, SL192, SL210, SL219, SL280, and SL339) and were subjected to the following polishing process. Reads were trimmed with fastp80, and, where possible, taxonomically filtered with Kraken281 to retain Pseudomonas (taxid 286). Genomes were assembled using SPAdes82, polished once with Pilon83 using bwa-mem2 read-mapping. Contigs <1 kb were discarded, and assembly metrics were assessed with QUAST84. Once all assemblies were checked for quality, annotations were performed using the RAST tool kit (RASTtk)85. Assembly quality metrics for all strains are included in Supplemental Data File 2.
SNP analysis was performed via kSNP4 to construct a maximum likelihood phylogenetic tree, and iTOL86 was utilized for editing the tree. Gene presence was determined by BLASTP (see Supplemental Data File 3 for sequences used) with a threshold of >85% identity and >90% query cover. ST typing was performed using the PubMLST Pseudomonas aeruginosa typing database39. Allelic matches are provided in Table 2. Serotypes were determined using the PAst v1 in silico serotyping tool40. To determine the presence of fleQ or lasR mutations, multiple sequence alignment was performed via Mafft. P. aeruginosa isolates from the PATRIC database stratified by infection type (UTI, wound, and lung) genome accession number can be found in Supplemental Data File 4.
Statistical Analysis
Competitive median (CM) was computed as 1/(median CFU of isolates with the gene of interest)/(median CFU of isolates without the gene of interest). Statistical analyses were performed using GraphPad Prism (v10.4). Variables were screened for correlation via Pearson’s correlation. Shapiro-Wilk normality tests and D’Agostino & Pearson omnibus normality tests were used to determine normality of data, and variances were compared using the F test. Group comparisons were made using unpaired two-tailed t-tests or Mann-Whitney U tests (95% CI) based on data distribution. When comparing two categorical variables, Fisher’s exact test was used. P values < 0.05 were considered statistically significant.
Ethics
This study was approved by the University of South Alabama Institutional Review Board (IRB, protocol #2178590). Human derived bacterial strains were obtained in accordance with the approved protocol. All bacterial isolates were coded prior to entering the research laboratory to ensure patient confidentiality. All animal protocols were approved by the Institutional Animal Care and Use Committee (IACUC, protocol #2006187) at the University of South Alabama College of Medicine.
ACKNOWLEDGMENTS
We would like to thank the University of South Alabama Frederick P. Whiddon College of Medicine for the start-up funds to support this project in the laboratory of Dr. Allyson Shea. We would also like to thank Teresa Barnett for the collection of bacterial isolates.
SUPPLEMENTAL MATERIAL
Supplemental Figure 1. UTI pathogen distribution in Mobile, Alabama. Retrospective chart review was conducted to identify the top UTI causative agents across the USA healthcare system from June 1, 2024 to May 31, 2025. Positive UTI cases were identified via urine culture (n=6929). Organisms <1% are not shown.
Supplemental Figure 2. O11 serotype is correlated with extensive drug-resistance (XDR). Prevalence of extensive drug-resistance (XDR, shown in black) in strains with the O11 serotype versus all other serotypes. Fisher’s exact test was used to determine statistical significance (*P<0.05).
Supplemental Figure 3. Clinical P. aeruginosa strains show a high number of antibiotic resistance genes. The genomes of the 55 clinical P. aeruginosa isolates were blasted for antibiotic resistance genes in the Comprehensive Antibiotic Resistance Database (CARD). Perfect hits (dark blue) have all amino acids matching the CARD sequence, strict (light blue) hits indicate very close matches with slight variations in the sequence. White indicates the gene is absent.
Supplemental Figure 4. Correlation matrix of genotypic data to patient variables. Gene presence or absence was determined by BLAST and correlated to patient variables. Gene presence or absence and patient variables were tested for correlation in GraphPad Prism with Spearman r. Red indicates a positive correlation, and blue indicates a negative correlation. P values were determined with 95% confidence (*P<0.05, **P<0.005, ***P<0.0005, ****P<0.00005). Blank or excluded comparisons are indicated with an X through the cell.
Supplemental Figure 5. Growth in human urine is affected by virulence gene presence and O-antigen serotype. (A) Dot plot with growth in human urine in strains with (red) and without (blue) genes parS, armR, and toxA. Bar height indicates median AUC. (B-C) Plots of growth in human urine (AUC) in serotypes (B) O5 and (C) O11 (red) versus all other serotypes (blue). Each dot represents a clinical isolate. The median of each group is indicated with a black line. Statistical significance was determined via Mann-Whitney U test (*P<0.05).
Supplemental Figure 6. Correlation matrix of phenotypic data to patient variables. The 55 P. aeruginosa clinical isolates were subjected to Chrome Azurol S (CAS), blood agar, biofilm, and growth curve assays to obtain phenotypic data. Patient variables were obtained and tested for correlation with phenotypic data in GraphPad Prism with Spearman r. Red indicates a positive correlation, and blue indicates a negative correlation. P values were determined with 95% confidence (*P<0.05, **P<0.005, ***P<0.0005).
Supplemental Figure 7. Patients with strains of O6 serotype have higher levels of blood in urine. Urinalysis results from USA Hospital were obtained for the UTI urine samples of origin for the 55 clinical isolates. A plot is shown of blood cells in urine specimens containing P. aeruginosa strains of serotype O6 versus other serotypes. Statistical significance was determined via Mann-Whitney U test (*P<0.05).
Supplemental Figure 8. Murine urine CFU burden of P. aeruginosa clinical and type strains. Urine was plated every 24 hours to assess bacterial load. Colored symbols indicate the mean of three technical replicates for each individual mouse (n=5).
Supplemental Data File 1: Minimum inhibitory concentration (MIC) ranges used for antimicrobial resistance determinations.
Supplemental Data File 2: Assembly quality metrics for the 55 P. aeruginosa urinary isolates.
Supplemental Data File 3: Protein sequences used for BLAST analyses.
Supplemental Data File 4: Strains from the PATRIC database used for comparison across isolation sources.