Non-redundant transposon mutant libraries containing ΔnasT, ΔcheB, ΔkdpD, and ΔaauS that were developed using PA14 (PA14NR) (Liberati et al., 2006) were obtained from Harvard Medical School. All strains were grown on cetrimide agar (CA) (Sigma-Aldrich, St. Louis, MO, USA) plates for 16–18 h at 37 °C. Bacterial cultures were grown in Luria-Bertani (LB) medium (Becton Dickinson, Franklin Lakes, NJ, USA) under orbital shaking culture conditions at 200 rpm. Gentamicin (15 μg/mL) was used to grow the mutants, whereas 50 µg/mL carbenicillin (Sigma-Aldrich) was used to grow the strains harboring empty pUCP18 or its derivative plasmids. The mutant complementation strains were grown in the presence of 15 μg/mL gentamicin and 50 μg/mL carbenicillin. Bacterial growth was monitored by measuring the optical density at 600 nm [OD_600nm = 1.0 = 1.0 × 10⁹ colony-forming units (CFU/mL)] using a Gen 5 BioTek microplate reader (BioTek, Winooski, VT, USA). Bacterial cells were harvested using 4000 rpm (2701 ×g) centrifugation at 4 °C. They were washed once using phosphate buffered saline (PBS, pH 7.2) in all experiments, unless otherwise stated. The strains and plasmids, as well as primers used in the study are listed in Supplementary Tables S1, S2 , respectively.

To complement the transposon mutants ΔkdpD and ΔaauS, the corresponding genes (kdpD and aauS) were PCR-amplified from P. aeruginosa PA14 genomic DNA using specific primers ( Table S2 ). The column purified PCR-amplified DNA fragments, and the pUCP18 vector were restriction endonuclease digested with EcoRI and HindIII. The EcoRI/HindIII-digested kdpD and aauS genes and pUCP18 vector DNA were resolved on an agarose gel. This was followed by gel elution using a Cosmo Genentech gel elution kit (Seoul, Korea). The gel-eluted kdpD and aauS were cloned into EcoRI/HindIII sites of the pUCP18 vector to yield the recombinant plasmids pUCP18kdpD and pUCP18aauS, respectively. These plasmids were transformed into Escherichia coli DH5α, and clones were selected on LB agar plates containing 50 μg/mL carbenicillin. The plasmids were isolated using a Cosmo Genentech plasmid isolation kit and DNA was sequenced to verify the accuracy of the nucleotide sequences of the genes in pUCP18kdpD and pUCP18aauS. The empty control plasmid, pUCP18, was transformed into wild-type (WT), ΔkdpD and ΔaauS of P. aeruginosa PA14. The WT strain possessing pUCP18 (WT::pUCP18) was selected on LB agar plates containing 50 μg/mL carbenicillin. Furthermore, the mutant strains (ΔkdpD and ΔaauS) possessing the empty vector (ΔkdpD::pUCP18 and ΔaauS::pUCP18) were selected on LB-agar plates containing 15 μg/mL gentamicin and 50 μg/mL carbenicillin. To obtain the knockout-complemented strains (ΔkdpD::pUCP18kdpD and ΔaauS::pUCP18aauS), the recombinant plasmids pUCP18kdpD and pUCP18aauS were transformed into ΔkdpD and ΔaauS knockouts and selected on LB plates containing 50 μg/mL carbenicillin and 15 μg/mL gentamicin.

Biofilm quantification through crystal violet (CV) staining was performed as previously described (Chen et al., 2018). Briefly, 16 h cultures of WT, mutant, and mutant-complemented strains were grown in Muller Hinton broth (MHB) (Becton Dickinson) at 37 °C and diluted to OD₆₀₀ = 1 × 10⁸ CFU/mL in fresh MHB media. An aliquot of 100 μL bacterial suspension was added to a flat-bottom 96-well polystyrene microtiter plate. After 24 or 48 h incubation at 37 °C, unattached cells were removed using two washes of PBS (pH 7.2). Next, 100 μL of 99% methanol was added to each well for 15 min, followed by methanol removal and drying of the plates at room temperature (25±1 °C) to perform biofilm fixation. Then, 100 μL of 0.04% CV was added for 20 min to stain the biofilm. Samples were washed with sterile PBS three consecutive times to remove additional colorants. Finally, 33% acetic acid was used to solubilize the biofilm-bound CV. Absorbance was measured at 630 nm using a Gen5 BioTek microplate reader (BioTek Instruments, Inc., Agilent Technologies, CA, USA). pqsA is an important biofilm producing gene so, ΔpqsA used as a negative control in the current experiment.

Bacteria cultures were grown in MHB at 37 °C for 16–18 h and diluted to OD_600nm = 1 (1×10⁸ CFU/mL) in fresh MHB media. Then, 500 μL bacterial suspension was added into an 8-well chambered cover glass (Lab-Tek II 1.5 Borosilicate glass, Sigma-Aldrich) and incubated at 37 °C under static culture conditions. The culture medium was changed after 24 h to remove non-attached bacterial cells. After 24 or 48 h of incubation, the biofilms were gently washed with saline [0.85% (w/v) sodium chloride (NaCl)]. Final concentrations of 30 μM propidium iodide (PI), and 5 μM SYTO9 (Sigma-Aldrich) were used for biofilm staining (Chaurasia et al., 2016). The chambered cover glass was incubated in the dark at room temperature for 20 min. The stained biofilms were washed with PBS prior to confocal microscopy. Biofilms were examined using confocal laser scanning microscopy (CLSM, Zeiss, Oberkochen, Germany) at 20× or 63× magnifications. Image stacks were collected from 20 random points on each biofilm to obtain accurate mean values for biofilm thickness measurement.

Swarming motility was assessed as previously described, with slight modifications (Ha et al., 2014a). M8-supplemented swarming medium was used to prepare the swarm plates. Briefly, 6 g agar was added to 800 mL water (final concentration = 0.6%) and autoclaved to obtain a sterile, homogeneous agar suspension. A 5× M8 solution was prepared using 30 g Na₂HPO₄, 15 g KH₂PO₄, and 2.5 g NaCl in water by adjusting the final volume to 1 L, autoclaving, and finally adding a 5× M8 solution (200 mL) to the melted agar. Next, 20% glucose (final concentration = 0.2%), 20% casamino acids (final concentration = 0.5%), and 1 M MgSO₄ (final concentration = 1 mM) were added to molten agar. The agar medium was mixed gently and cooled before being poured into Petri dishes (~25 mL/plate) and solidified at room temperature. The center of each plate was inoculated with 2.5 μL of OD_600nm 1 equivalent overnight bacterial culture. Plates were incubated upright at 37 °C for 16–24 h and observed for the swarming phenotype. The sensor kinase fleS of the fleSR TCS was used as a negative control in the current experiment, as ΔfleS alters swarming motility in P. aeruginosa PA14 (Kollaran et al., 2019).

Swimming motility was assessed using 0.3% LB agar as previously described, with slight modifications (Déziel et al., 2001). First, 25 mL media plates were solidified at room temperature for 1.5 h. The cell number was adjusted based on the OD_600nm value. Then, 100 μL of OD_600nm 1 bacterial culture was transferred into a 1.5 mL microcentrifuge tube. To avoid the production of swarming and twitching phenotypes, a sterile toothpick was dipped into the OD_600nm 1 equivalent culture and stabilized in the middle of the agar layer without touching the top or bottom of the Petri plate. The plates were incubated upright for 20 h at 30 °C. The phenotype was observed, and the diameters of the swim zones were measured. The hook-associated gene, flgK, is essential for flagellar assembly. Additionally, its mutant showed reduced swimming motility (Ha et al., 2014b). Therefore, in the current experiment ΔflgK was used as a negative control for the no-swimming phenotypes.

Twitching motility was tested as previously described (Déziel et al., 2001). Twitching motility was assessed on LB plates containing 1% (w/v) agar. A sterile toothpick was dipped into the OD_600nm 1 equivalent bacterial culture. The toothpick touched the bottom of the Petri plate to move into the interstitial space between the basal surfaces of the agar. The plates were subsequently incubated at 37 °C for 48 h. The agar was gently removed from the Petri plate using a spatula to measure the twitch zone. The twitch zone was stained with 2 mL 0.1% (w/v) CV in water for 10 min. The CV was removed, and the plates were washed with water and allowed to air-dry at room temperature. The twitch zone diameters were measured and recorded. As the role of the TCS PilS/PilR in the regulation of type IV pilus expression has been demonstrated (Kilmury and Burrows, 2018a), we used ΔpilR as a negative control.

The WT, mutant, and complementation strains were grown in M9 minimal medium for 10 h at 37 °C (Egli, 2015). For identification of biofilm genes bacteria was grown in 8-well chambered cover glass for 10, 24 and 48 h. Planktonic cells were discarded, and biofilm population was used for RNA isolation (Bisht et al., 2021). Total RNA was isolated using the Qiagen RNeasy kit (Hilden, Germany) according to the manufacturer’s protocol. Total RNA (1 µg) was treated with RNase-free DNase I (1 U) (amplification grade; Sigma-Aldrich) at room temperature for 15 min. The DNase I reaction was stopped using stop-buffer followed by inactivation of DNase at 70 °C for 10 min. To prepare cDNA from the isolated DNA-free mRNA, a random hexamer premix (RNA-to-cDNA EcoDryTM premix; Takara Bio, Kusatsu, Japan) was used. The qRT-PCR was performed using 55 °C annealing temperature and SYBR Green Supermix (Bio-Rad, Hercules, CA, USA). Here, rho was used as a housekeeping gene. The expression of kdpD and aauS was normalized to calculate the relative gene expression analyzed using the 2^−ΔΔCT method (Schmittgen and Livak, 2008).

To examine the roles of kdpD and aauS in P. aeruginosa virulence, we performed invasion assays as previously described (Kim et al., 2019). HeLa cells were grown in Dulbecco’s modified Eagle’s medium (DMEM) with 10% fetal bovine serum (FBS) in a 5% CO₂ humidified incubator at 37 °C. The cells were seeded in 6-well plates (5.0×10⁵ cells/well) and incubated for 24 h. PBS-washed bacterial cells were placed in invasion medium (DMEM without FBS) for 2 h before infection. Then, WT and mutant strains at a multiplicity of infection (MOI) of 10 were used to infect HeLa cells for 120 min. Extracellular bacterial cells were killed through 60-min gentamicin (100 µg/mL) treatment. Residual gentamicin was removed by washing the cells twice with PBS. The infected cells were treated with 0.1% Triton-X100 to lyse the HeLa cells and retrieve intracellular bacteria, which were then serially diluted in PBS to plate 100 µL diluted cell suspension on cetrimide agar plates for CFU enumeration.

The live/dead assay was performed as previously described (Mittal et al., 2014). Briefly, DMEM with 10% FBS was used to grow HeLa cells in a 5% CO₂ humidified incubator at 37 °C for 48 h. HeLa cells were seeded into 8-well chamber slides (5.0×10⁵ cells/well) and incubated for 24 h. After two PBS washes, HeLa cells were infected with bacteria at an MOI of 10 for 120 min. Uninfected cells were used as negative (untreated) controls. After incubation, the infected HeLa cells were washed with PBS and a live/dead viability kit was used for staining. Fluorescently stained bacterial cells were observed using CLSM (Zeiss, Oberkochen, Germany) and images were captured.

The overnight grown cultures of P. aeruginosa strains were suspended in LB for 6 h. Bacterial cells were collected through 4000 rpm centrifugation (2701 ×g) at 4 °C and washed with PBS (pH 7.2). The number of bacterial cells was maintained by adjusting the optical density (OD_{600 nm} = 1) in PBS. G. mellonella larvae were stored for 24 h before infection (Ménard et al., 2021). Then, the weight and length of each worm were measured. Ten waxworms of similar size (200 ± 2 mg, 2 ± 0.2 cm) were chosen for each group, including the carrier (PBS) control. Waxworms were injected with 10 bacterial cells (20 µL) in the last posterior leg using a 0.3 mL syringe (Becton Dickinson) (Desbois and Coote, 2011; Hill et al., 2014; Imdad et al., 2018). Similarly, 20 µL PBS was injected into the worms in the control group. After injection, worms were incubated at 27 °C and observed at different time points. We applied a health scoring index wherein each worm was given a score based on movement, melaninization, cocoon formation, and survival compared to the untreated control group (Champion et al., 2018). The experiment was terminated after 72 h and CFUs were counted using CA plates.

The relevance of kdpD, aauS, nasT, and cheB in P. aeruginosa virulence is yet to be well characterized (Francis et al., 2017). Therefore, we initially tested their roles in virulence using a G. mellonella infection model. The hemocytes present in G. mellonella larvae share functional similarities with mammalian phagocytic cells (Browne et al., 2013). Therefore, G. mellonella larval infection models have been used to identify virulence-related genes in various human pathogens (Jander et al., 2000; Champion et al., 2009; Lebreton et al., 2009; Peleg et al., 2009; Champion et al., 2010). rsmA, a virulence gene encoding the carbon storage and a pyocyanin production regulator, was used as a positive control to determine the potential roles of uncharacterized regulators. Mutant strains lacking rsmA, kdpD, aauS, nasT, cheB, or the WT were injected into the left posterior legs of G. mellonella larvae (n = 10) to assess their roles in the pathogenic potential of P. aeruginosa PA14. The results showed only 10% survival in WT strains 72 h post-infection, whereas the ΔrsmA showed 90% survival compared to the PBS control (100% survival) ( Figure 1 ; Supplementary Figure S1 ). However, a delay in mortality was observed for ΔkdpD and ΔaauS, with 60% and 30% survival, respectively. The comparative growth profile of ΔkdpD and ΔaauS knockouts with the WT strain was monitored for physiological fitness to confirm that the difference in their infection potential was not due to their compromised growth proficiency. The growth of the WT, ΔkdpD, and ΔaauS strains was identical ( Supplementary Figure S2 ). Furthermore, the ΔnasT and ΔcheB strains showed earlier mortality than the WT, with only 20% survival 72 h post-infection. These results demonstrate that kdpD and aauS contribute to virulence. Thus, we further investigated their relevance to the virulence potential of P. aeruginosa.

KdpD/KdpE is an ‘adaptive regulator’ TCS involved in the virulence and intracellular survival of various pathogenic bacteria (Freeman et al., 2013). To elucidate the roles of kdpD and aauS in biofilm formation, we used a CV staining assay, followed by CLSM confirmation and image analysis to determine the quantity and thickness of the biofilm, respectively. Firstly, strains ΔkdpD and ΔaauS were tested against the WT (positive control) and ΔpqsA (negative control). After 24 and 48 h, ΔpqsA showed considerably poor biofilm formation compared to the WT. At 24 h, ΔkdpD and ΔaauS showed 69.1% and 54.9% reduction in biofilm formation, respectively ( Supplementary Figure S3A ). Similarly, we observed 69.7% and 72.8% inhibition of biofilm formation after 48 h in ΔkdpD and ΔaauS, respectively ( Supplementary Figure S3B ). Furthermore, knockouts were complemented with their corresponding genes in the pUCP18 vector to confirm the roles of kdpD and aauS in biofilm formation. Empty plasmid controls (pUCP18) were used for WT strains and knockouts. At 24 h, knockout strains with the empty plasmids ΔkdpD::pUCP18 and ΔaauS::pUCP18 showed 38.7% and 29.0% reduction in biofilm formation, respectively compared to WT::pUCP18 ( Figures 2A, B ). Similarly, ΔkdpD::pUCP18 and ΔaauS::pUCP18 showed 62.7% and 58.7% biofilm reduction, respectively at 48 h. The complemented strains ΔkdpD::pUCP18kdpD, and ΔaauS::pUCP18aauS showed biofilm formation similar to that of WT::pUCP18 ( Figures 2C, D ). These results indicate that kdpD and aauS contribute to biofilm formation. Furthermore, we investigated biofilm formation in the WT, ΔkdpD, ΔaauS, and complementation strains through CLSM using empty plasmid controls. At 24 and 48 h, ΔkdpD::pUCP18 and ΔaauS::pUCP18 showed reduced biofilm formation compared to WT::pUCP18, whereas their complemented strains ΔkdpD::pUCP18kdpD and ΔaauS::pUCP18aauS exhibited similar biofilm formation as WT::pUCP18 ( Figures 2E–H ). Based on these results, we concluded that kdpD and aauS play vital roles in the biofilm formation of P. aeruginosa.

Bacterial motility is an important virulence factor in host-pathogen interactions. Therefore, we investigated the relevance of kdpD and aauS in motility by analyzing the comparative swarming phenotypes of the WT with ΔkdpD, ΔaauS, and mutant-complemented strains. Our data revealed markedly reduced swarming (69.7%) in ΔfleS compared to the WT ( Figures 3A, B ). Contrastingly, ΔkdpD and ΔaauS showed 37.5% and 35.8% reduction in swarming motility, respectively compared to the WT ( Figures 3A, B ). These results indicate that kdpD and aauS contribute to swarming motility.

The swimming phenotype is also important for flagellar motility, which plays an important role in infection by various pathogenic bacteria (Horstmann et al., 2017; Corral et al., 2020). Therefore, the involvement of kdpD and aauS was investigated to assess their roles in swimming motility. Here, ΔflgK showed a reduction of 62.9% in swimming compared to WT, whereas ΔkdpD and ΔaauS did not ( Figures 3C, D ). These findings suggest that these two genes are not involved in swimming motility in P. aeruginosa. Furthermore, we compared the twitching phenotypes of ΔkdpD and ΔaauS knockouts with those of the WT strain. We observed 62.9% and 26.5% reductions in twitching activity in ΔpilR and ΔkdpD compared to the WT. However, no reduction was observed in the twitching phenotype of the ΔaauS mutant strain ( Figures 3E, F ).

We confirmed these results by investigating the motility of WT::pUCP18, ΔkdpD::pUCP18, ΔaauS::pUCP18, ΔkdpD::pUCP18kdpD, and ΔaauS::pUCP18aauS. In this study, ΔkdpD::pUCP18 and ΔaauS::pUCP18 showed reduced swarming compared to the empty vector control (WT::pUCP18). In contrast, their corresponding mutant-complemented strains ΔkdpD::pUCP18kdpD and ΔaauS::pUCP18aauS exhibited a swarming phenotype similar to that of WT::pUCP18 ( Figure 3G ). These results indicate that kdpD and aauS contributory roles into the swarming mobility of P. aeruginosa. Complementation of the sensor kinases aauS and kdpD further confirmed that these genes have no role in swimming motility ( Figure 3H ). aauS displayed no role in twitching, whereas kdpD played a considerable role in twitching mobility. Similarly, ΔkdpD::pUCP18 showed reduced twitching, and ΔkdpD::pUCP18kdpD restored the phenotype to WT::pUCP18 ( Figure 3I ). These results indicate that kdpD contributes to swarming and twitching motility, whereas aauS only affects the swarming motility of P. aeruginosa.

Well-known genes were initially considered to assess the relative gene expression in ΔkdpD and ΔaauS knockouts with the empty plasmids to elucidate the role of kdpD and aauS in biofilm formation, swarming, and twitching mobility. Many genes play important roles in P. aeruginosa biofilm formation. For example, lasA plays a key role in biofilm development (Thi et al., 2020) while the pelA markedly contributes to the pellicle matrix formation of biofilm in P. aeruginosa (Friedman and Kolter, 2004; Sakuragi and Kolter, 2007; Colvin et al., 2012; Wei and Ma, 2013). Additionally, cupA is a crucial fimbria gene that contributes to biofilm formation via adhesion (Vallet et al., 2001; Vallet-Gely et al., 2007). Therefore, lasA, pelA, and cupA expression was investigated in the WT with empty plasmids and knockout with the empty plasmid strains. We found reduction in lasA expression at 10 h by 62.5% and 39.1% in ΔkdpD::pUCP18 and ΔaauS::pUCP18, respectively ( Figure 4A ). Complemented strains ΔkdpD::pUCP18kdpD and ΔaauS::pUCP18aauS restored expression levels to the level of of WT::pUCP18 ( Figure 4A ). Furthermore, lasA expression was also found to be reduced at 24 and 48 h time points by 25.9% and 19.1% in ΔkdpD::pUCP18; and 22.0% and 26.0% in case of ΔaauS::pUCP18, respectively ( Figures 4B, C ). These results demonstrate that kdpD and aauS contribute to initial biofilm formation by controlling lasA expression. We found no significant reduction in pelA and cupA in both the ΔkdpD::pUCP18 and ΔaauS::pUCP18 strains at 10 and 24 h ( Supplementary Figures S4A–D ). However, at 48 h pelA showed slightly reduced expression of 13.0% and 31.0% in the ΔkdpD::pUCP18 and ΔaauS::pUCP18 strains, respectively ( Figure 4D ). Furthermore, cupA also showed reduced expression of 21.9% and 15.3% at 48 h in the ΔkdpD::pUCP18 and ΔaauS::pUCP18 strains, respectively ( Figure 4E ). Complemented strains ΔkdpD::pUCP18kdpD and ΔaauS::pUCP18aauS restored expression levels to the level of WT::pUCP18 ( Figure 4E ). Overall, these results indicate that kdpD and aauS contribute to biofilm formation by modulating the expression of lasA, pelA and cupA genes.

Many genes and transcription regulators involved in flagellar biosynthesis, assembly and regulation are relevant to motility (Dasgupta et al., 2003). For example, fleQ is an important transcription regulator in P. aeruginosa, contributing to transcription of many flagellar genes (Arora et al., 1997). Similarly, FleS/FleR is a crucial TCS that modulates flagellar motility and adhesion (Ritchings et al., 1995). FleN is an anti-activator of FleQ, which plays a vital role in retaining a single flagellum (Dasgupta et al., 2000). In addition, sigma factor σ54 encoded by rpoN contributes to the regulation of flagellar expression (Totten et al., 1990). Furthermore, flgM is an important flagellar gene whose interactions with the sigma factor fliA contribute to flagellar biogenesis (Frisk et al., 2002). We analyzed the expression of several flagellar genes and transcription regulators in ΔkdpD and ΔaauS strains to investigate kdpD and aauS contribution to motility. There were no significant differences in the expression levels of the major motility-regulated genes fleS, fleR, fleQ, flgK, fliC, fliD, flgD, and flgE ( Supplementary Figure S5 ). Therefore, a top-down approach was applied to comprehensively elucidate the mechanistic roles of kdpD and aauS in biofilm, motility, and pathogenesis. The expression of known transcription factors (TF) in ΔkdpD::pUCP18 and ΔaauS::pUCP18 compared to WT::pUCP18 was examined using q-RT-PCR. Accordingly, the transcriptional regulators (algR, cpxR, furR, rpoD, rpoN, rpoS, lasR, algU, anR, narL, and nbtR) that are directly or indirectly involved in biofilm formation, motility, and virulence gene expression have been assessed (Pusic et al., 2021). In this study, the expression levels of lasR was enhanced by 190% and 110% in ΔkdpD::pUCP18 and ΔaauS::pUCP18, respectively, compared to WT::pUCP18 ( Figure 4F ). Secondly, the anR expression was increased by 130% and 306% in ΔkdpD::pUCP18 and ΔaauS::pUCP18, respectively; compared to WT::pUCP18 ( Figure 4F ). anR and lasR contribute to biofilm formation and motility by controlling the Pseudomonas quinolone signal (PQS) quorum sensing (QS) system (Pusic et al., 2021). The las-QS system is essential for PQS-stimulated biofilm formation in P. aeruginosa PA14 (Christiaen et al., 2014). These results partially demonstrate that kdpD and aauS contribute to biofilm formation by controlling lasA expression through the lasR TF. PQS suppresses swarming motility in P. aeruginosa PA14 (Guo et al., 2014; García-Reyes et al., 2021). Therefore, we confirmed the expression levels of pqs in ΔkdpD::pUCP18 and ΔaauS::pUCP18. Our results showed increased pqsA, pqsB, pqsC, and pqsD expression (141%, 208%, 115% and 230%) in ΔkdpD::pUCP18, and (169%, 228%, 107% and 127%) in ΔaauS::pUCP18 strains ( Figure 4G ). Their expression was restored to that observed in WT::pUCP18 in the complemented (ΔkdpD::pUCP18kdpD and ΔaauS::pUCP18aauS) strains. Overall, these results indicate that kdpD and aauS contribute to swarming motility by altering the expression of the PQS quorum sensing system controlled by anR and lasR transcriptional regulators.

P. aeruginosa mainly invades mammalian epithelial cells via the type III secretion system (T3S) (Hernández-Padilla et al., 2017; Kroken et al., 2018). Therefore, we investigated the invasive activity of mutant strains with empty plasmids to evaluate the roles of kdpD and aauS in cell invasion. HeLa cells were infected with bacteria at an MOI of 10 for 120 min, and the number of intracellular bacterial cells was measured through CFU enumeration. Strains ΔkdpD::pUCP18 and ΔaauS::pUCP18 showed reduced invasion activity, with intracellular CFUs difference of 2.0×10³ and 2.3×10³ CFU/mL, respectively compared to WT::pUCP18. In contrast, the intracellular CFU of the WT::pUCP18 strain was 3.7×10³ CFU/mL ( Figure 5A ). Moreover, the complemented strains ΔkdpD::pUCP18kdpD and ΔaauS::pUCP18aauS restored the intracellular CFU count to that of WT::pUCP18.

The host-pathogen complex was stained with live/dead (green/red) reagents to further evaluate the roles of kdpD and aauS in HeLa cell infection. Uninfected HeLa cells mostly retained green fluorophores (Untreated, Figure 5B ). A high number of dead host cells was observed when WT::pUCP18 was added ( Figure 5B ). Moreover, a markedly increased number of live cells was observed when host cells were infected with ΔkdpD::pUCP18 or ΔaauS::pUCP18 ( Figure 5B ). Contrastingly, a considerable number of dead host cells was observed when the complemented strains ΔkdpD::pUCP18kdpD and ΔaauS::pUCP18aauS were used to infect HeLa cells. These results demonstrate the involvement of kdpD and aauS in host cell invasion and infection.

G. mellonella larvae were challenged with WT::pUCP18, ΔkdpD::pUCP18, ΔaauS::pUCP18, ΔkdpD::pUCP18kdpD, and ΔaauS::pUCP18aauS at 10 CFU/larvae to further analyze the roles of kdpD and aauS in virulence. The WT::pUCP18 bacterial load was designed to allow 10% survival 72 h post-infection. Then, survival percentages of the larvae and their health indices were recorded. Our results indicated delayed mortality, with increased survival in ΔkdpD::pUCP18 (60%) and ΔaauS::pUCP18 (40%) compared to WT::pUCP18, after 72 h. This is consistent with the results obtained in the initial screening ( Figure 1 ). We further investigated the infection properties of complemented ΔkdpD::pUCP18kdpD and ΔaauS::pUCP18aauS strains. The results showed a survival comparable to that of WT::pUCP18 cells ( Figure 6A ). We also recorded the health index of each strain based on movement, cocoon formation, melaninization, and survival. Our results showed greater movement, full cocoon formation with less melaninization, and increased survival in larvae infected with ΔkdpD::pUCP18 and ΔaauS::pUCP18 than in those infected with WT::pUCP18. However, larvae infected with ΔkdpD::pUCP18kdpD and ΔaauS::pUCP18aauSshowed less movement, greater melaninization, less cocoon formation, and decreased survival 72 h post-infection, demonstrating results similar to those of WT::pUCP18 ( Figures 6B, C ). Next, we determined the bacterial burden in the worms by measuring the number of bacteria in the infected larvae. Larvae infected with ΔkdpD::pUCP18 and ΔaauS::pUCP18 showed 1.7×10⁵ and 3.4×10⁵CFU/mL, respectively. These were lower than the 9.3×10⁵ CFU/mL observed in the WT::pUCP18 strain ( Figure 6D ). However, strains ΔkdpD::pUCP18kdpD and ΔaauS::pUCP18aauS showed no significant reduction in CFU compared to WT::pUCP18 ( Figure 6D ). These results were consistent with the survival and health index results, indicating that kdpD and aauS play important roles in G. mellonella infection.