The type III secretion system facilitates systemic infections of Pseudomonas aeruginosa in the clinic

ABSTRACT Pseudomonas aeruginosa is a major opportunistic pathogen that can cause severe systemic infections. Here, we found that the type III secretion system effector ExoU was the main determinant of pathogenicity of a highly virulent clinical isolate strain, BSI_S5, of P. aeruginosa, which caused severe bloodstream infections and belongs to ST463/O4. Deletion of exoU showed significantly attenuated cytotoxicity and virulence in vivo, while deletion of two other unique genes from strain BSI_S5, chr_1696 and chr_4238, which encodes type I secretion system permease/ATPase or polysaccharide biosynthesis protein, respectively, caused no apparent loss of toxicity compared to the parent strain, and complementation of the ΔexoU mutant restored its virulence. The presence of ExoU-induced lung lesions and the mucosal damage of gallbladder in mice, and the replication of the ΔexoU mutant was hindered in vivo compared with the parent strain. We show here for the first time that a clinical isolate with intact exoU harbored a mutation of T→G substitution in the adjacent specific Pseudomonas chaperone for ExoU (spcU), which caused phenylalanine to valine change at amino acid 94 of SpcU, resulting in significant loss of cytotoxicity. This finding suggests that intact exoU is not sufficient for cytotoxicity, but a functional downstream SpcU is required for ExoU secretion and cytotoxicity. Finally, we found that BSI_S5 harbored two types of insertion sequences adjacent to exoU, suggesting the potential of exoU to be transferred to other strains. Our study opens an avenue for further mechanistic study of P. aeruginosa systemic infections regarding its high pathogenicity and clinical spread. IMPORTANCE The identification of decisive virulence-associated genes in highly pathogenic P. aeruginosa isolates in the clinic is essential for diagnosis and the start of appropriate treatment. Over the past decades, P. aeruginosa ST463 has spread rapidly in East China and is highly resistant to β-lactams. Given the poor clinical outcome caused by this phenotype, detailed information regarding its decisive virulence genes and factors affecting virulence expression needs to be deciphered. Here, we demonstrate that the T3SS effector ExoU has toxic effects on mammalian cells and is required for virulence in the murine bloodstream infection model. Moreover, a functional downstream SpcU is required for ExoU secretion and cytotoxicity. This work highlights the potential role of ExoU in the pathogenesis of disease and provides a new perspective for further research on the development of new antimicrobials with antivirulence ability.


Virulence of P. aeruginosa isolates in chick embryo model and cytotoxicity assay
The chicken embryo could be served as a reliable, inexpensive, and easy setup for assessing bacterial virulence, and it has been used for different bacterial spe cies including Listeria monocytogenes, Salmonella Typhimurium, Staphylococcus aureus, Escherichia coli, Enterococcus cecorum, and P. aeruginosa.(20)(21)(22)(23)(24)(25)(26)(27).We used the chicken embryo lethality assay to assess the pathogenic potential of different clinical strains of P. aeruginosa with PAO1 as a reference strain.As shown in Fig. 1a, the survival rates at 18th hour post-infection (hpi) were lower in embryos infected with strain BSI_S5 (60%), BSI_S3 (80%), and S1 and COPD_S2 (both 90%) than those infected with AECOPD_S4 and PAO1 (both 100% survival).Among the P. aeruginosa clinical strains, the survival rate on the 32nd hpi was the lowest for S4 and S5 (0%), demonstrating the highest embryo mortality and shortest survival.Finally, infected embryos all died at 44 hpi when a dose of 10 5 CFU of different strains of P. aeruginosa was applied.
We also compared the morphological features of the infected chicken embryos, and the results showed that the infected embryos were identified by the absence of spontaneous movement, usually associated with hemorrhage, and the small black dot at the other end of the air chamber, loss and blackening of vascular architecture, and abnormal morphology.The loss and blackening of vascular architecture were the most severe for embryos infected by BSI_S5 on the 32nd hpi.
To further evaluate the cytotoxicity of the clinical strains of P. aeruginosa, we used a commonly adopted cell infection model with THP-1 cells.The infected THP-1 cells had 28.09% ± 4.06% viability at 3.5 hpi with BSI_S5, which was significantly higher than all the other groups (P < 0.0001).The value for PAO1, S1 or AECOPD_S4 was 73.22% ± 2.28%, 68.79% ± 3.14%, and 77.56% ± 0.13%, respectively, and these values were significantly higher than uninfected cells [uninfected control (CT) or phosphatebuffered saline (PBS)] (P < 0.0001).The cell viability for THP-1 cells infected with COPD_S2 and BSI_S3 was 89.00% ± 0.97% and 91.68% ± 1.20%, respectively, demonstrating their low cytotoxicity (Fig. 1b).Also, we used the annexin V-FITC/PI (fluorescein isothiocyanate/propidium iodide) to determine if there was externalization of phosphatidylserine residues on the outer plasma membrane of apoptotic cells.THP-1 cells infected with BSI_S5 showed the highest proportion of double staining by annexin V-FITC/PI, which indicates apoptotic and necrotic cell death (Fig. 1c).Taken together, these results demonstrated that S5 was the most virulent among the strains.S5 was isolated from the blood sample of a 53-year-old female patient hospitalized with acute myeloid leukemia complicated with bilateral lung and biliary infections.It was isolated as the causative pathogen of systemic infections (it also existed in urine and sputum).The clinical background and bacterial isolates used are listed in Table S1 through S2.

Features of P. aeruginosa genomes
To identify the molecular basis for the high-virulence strain BSI_S5, a hybrid assembly using PacBio sequencing as well as Illumina sequencing was performed to determine the complete genome sequence of the strain .The assembled genome of BSI_S5 has a single circular chromosome of 6,920,497 bp with an average GC content of 65.86% (Fig. 2a), which is larger than complete genomes of the reference strain P. aeruginosa PAO1 (6,268,251 bp) and other clinical isolates such as S1, BSI_S3, and AECOPD_S4, but smaller than COPD_S2 (7,181,513 bp) (Table 1).A sum of 6,252 genes were predicted from the S5 genome by the NCBI Prokaryotic Genome Annotation Pipeline server, which included 6,186 coding sequences and 67 tRNA genes and 16 rRNA operons and 79 ncRNA genes.
PA14, which was recognized as a virulent strain (28), was included as a control in the comparative genomics analysis.Comparison of the annotated genes in the P. aeruginosa PAO1 and PA14 genomes with those clinical P. aeruginosa genomes revealed an extensive conservation of a set of genes that are shared by all of the strains ( Fig. 2b and c).A total of 5,163 genes are conserved across all seven genomes analyzed, which are recognized as the P. aeruginosa core genome (60.62% of the whole genome).The genome for each strain carries a relatively modest number of unique sequences, where unique gene family numbers are 329, 49, 114, 432, 407, 49, and 369 for strain BSI_S5, AECOPD_S4, BSI_S3, COPD_S2, S1, PAO1, and PA14, respectively, confirming that P. aeruginosa harbors a large amount of strainspecific variations in protein-coding genes.Besides, there were three plasmids in S1 and two in BSI_S5, which can transfer horizontally between bacterial cells, within and between communities of the same or different species and play a crucial role in bacterial ecology and evolution (29).

Virulence Factor Database analysis
The analysis and comparison of genomes showed a diversity of speciesspecific genes between strains.The distribution of virulence-associated genes was further analyzed among the core and accessory genome of these strains.Previous studies have identified  (mucoid exopolysaccharide), quorum sensing], and nutritional/metabolic factor (i.e., pyochelin, pyocyanin and pyoverdine).As for the most virulent strain S5 in this study, 442 genes related to the virulence were identified, and their locations are shown in Fig. 3. Apart from the general recognized virulence factors, some of these genes were related to antimicrobial activity/competitive advantages, invasion, post-translation modification, regulation, and stress survival.Three genes were unique to S5 among the six strains of P. aeruginosa analyzed in this study, which are located at chromosomes 1696, 1770, and 4238, respectively (Table 2 ).
As for the chr_1696 (2,148 bp, type I secretion system permease/ATPase), BLAST alignment analysis showed its nucleotide sequence has 100% identity with the sequence in different strains of P. aeruginosa.Proteins encoded by chr_1696 show a high level of amino acid homology to cyclolysin secretion ATP-binding protein.
The sequence of chr_1770 (2,064 bp) showed 100% identity with exoU, which encodes a major virulence determinant of P. aeruginosa-T3SS effector cytotoxin ExoU.T3SS also injects other potent cytotoxins, including ExoS, ExoT, or ExoY, into eukaryotic cells.The production of each of the different enzymes determines a distinct host tissue injury, where ExoU is the one that has a greater impact on bacterial virulence (13) and has been associated with increased morbidity and mortality in patients with pneumonia and bacteremia (12,13).The nucleotide sequence of chr_4238 (1,035 bp) showed 100% identity with the sequence in different strains of P. aeruginosa.The protein encoded by chr_4238 showed a high level of amino acid homology to type VIII capsular polysaccharide synthesis protein Cap8E or capsular polysaccharide biosynthesis protein Cps4J.Capsular polysaccharides are one of the major contributors to virulence of various microorganisms, as the presence of capsule enables these bacteria to escape from detection and clearance by the host immunity (30).

exoU is critical for increased virulence in P. aeruginosa-induced infection
To identify which of the above three genes might play a role in determining the high virulence of P. aeruginosa BSI_S5, we first performed deletion mutant construction (Fig. S1) followed by complementation studies of the above-mentioned candidate genes.To determine if the deleted gene affected the growth of BSI_S5, their ability to grow in Luria-Bertani (LB) medium was tested by measuring the optical density at 600 nm.Results showed that there was no significant difference in growth curves between the parent strain and the deletion mutants (Fig. S2).
Next, we adopted a widely used cytotoxicity model, which is based on evaluating the toxicity of the pathogen to THP-1 macrophages, to determine the virulence of the deletion mutants.Trypan blue staining showed the deletion of chr_1696 and chr_4238 caused no apparent loss of cytotoxicity, while deletion of exoU showed significantly decreased toxicity compared to the parent strain (P < 0.0001).In contrast, ∆exoU deletion mutation caused a dramatic loss of cytotoxicity with no significant difference between CT (P > 0.05), indicating severe attenuation of cytotoxicity for the exoU mutant (Fig. 4a).
In addition, we used lactate dehydrogenase (LDH) assay to further confirm the cytotoxicity for the parent strain BSI_S5 and the deletion mutants.LDH assay is a common method for evaluating cytotoxicity, which is based on determining the activity of cytoplasmic enzymes released by damaged cells when cells undergo apoptosis, necrosis, and other forms of cellular damage (31).As shown in Fig. 4b, deletion of exoU showed attenuated cytotoxicity in the LDH assay, whose value was 2.99% ± 0.08% at 2.5 hPi, whereas the parent strain BSI_S5 induced 43.22% ± 0.46%.The polysaccharide biosynthesis protein mutant (∆4238) showed a decreased LDH value compared to the parent strain BSI_S5 (P < 0.0001), while the attenuation of ∆4238 mutant was less obvious when the infection time was increased to 4.5 hours (its virulence was not different from the parent strain at 4.5 hours) (P > 0.05) (Fig. S3a).However, no decrease in cytotoxicity was observed for the ∆1696 mutant even when infection time increased to 4.5 hours (Fig. S3a), suggesting the type I secretion system permease/ATPase does not contribute to the cytotoxicity.

aeruginosa virulence in mouse systemic infection in vivo
To further verify the role of ExoU during infection, P. aeruginosa BSI_S5 [parent strain, marked as wild type (WT)] and ΔexoU mutant were employed in a bloodstream infection utilizing a mouse model via tail vein.The results showed that the ΔexoU mutant-infec ted group had a 63.6% survival rate, versus only 14.3% (P < 0.05) for parent BSI_S5 over a period of 15-day observation.When the exoU gene was complemented in trans [pK∆exoU (exoU)], the survival rate of mice was comparable to that of WT (Fig. 5a).We also examined the fate of P. aeruginosa during the bloodstream infection to allow for consistent and reproducible delivery of bacteria directly into the bloodstream.Enumera tion of bacterial counts from infected organs was conducted from mice infected with parent strain and ΔexoU.When each mouse was infected with a dose of 5 × 10 6 CFU by tail-vein injection, P. aeruginosa was detected in the lungs, spleen, liver (Fig. 5b through  d), and a minor proportion of blood (data not shown), but no P. aeruginosa spp.were detected in gallbladder (GB) and feces (data not shown) at 22 hpi.Bacterial recovery in the lung and spleen reached significantly higher numbers in the parent strain infection group than those for ΔexoU (P < 0.05), suggesting that ΔexoU replication was hindered in vivo compared with the parent strain.When a high dose of 4 × 10 7 CFU was applied, significant difference of bacterial recovery of parent strain was observed at 22 hpi in lungs, spleen, and liver, and a higher amount was detected in blood (Fig. 5e through h).Besides, a dramatic expansion of the bacterial population in the GB was observed for the majority (four of seven) of parent strain-infected group but only one of seven for ΔexoU-infected group (Figure 5i).Also, the dramatic expansion of the bacterial population was detected in feces for parent strain-infected mouse, which was significantly higher than ΔexoU-infected mouse (Fig. 5j).
We examined GBs of infected animals for histological evidence of injury.The wall of GBs from mock-infected mouse (control, PBS) consists of the mucosa, lamina propria, an irregular muscular layer, perimuscular connective tissue, and serosa (or visceral peritoneum) (32)(33)(34).The connective tissue along the hepatic surface is continuous with the interlobular connective tissue of the liver (33).In contrast, the mucosa of GBs from parent strain-infected mice displayed a general distortion of the cellular architec ture characterized by disrupted lamina propria and damaged columnar epithelial cells.However, in contrast to the mucosal damage observed with parent strain infection, no damage to the GB epithelium was observed for ΔexoU-infected mice (Fig. 5k).We then evaluated pulmonary lesions by histopathological analyses.Massive infiltration of inflammatory cells, destruction of alveolar wall structure, and detachment of epithelial cells were observed in the parent strain-infected group, whereas ΔexoU induced less detachment of epithelial cells compared with mice infected with parent strain.
The majority of previous studies mainly focused on the correlation between virulence and the exoU gene (10)(11)(12)(13)35).For example, P. aeruginosa strains isolated from hospi tal-acquired pneumonia patients were highly virulent if the isolate harbored the exoU gene compared with isolates that did not (36).Although exoU gene was known to be a major contributor to potential virulence of P. aeruginosa in terms of cytotoxicity (37,38), in this study, we used comparative genomic analysis to identify the three unique genes, chr_1696, chr 4238, and exoU.By gene deletion and virulence screens, we finally confirmed that the exoU gene was the determinant affecting the virulence of P. aeruginosa BSI_S5 at the molecular level by gene deletion studies followed by virulence testing using cell and animal models, beyond a correlation analysis.

Prevalence of exoU gene in clinical isolates of P. aeruginosa
Since our gene knockout and virulence analysis showed that only exoU was the main driving force that caused the high virulence of BSI_S5, we wanted to know if this gene is commonly present in clinical isolates of P. aeruginosa from different sources.To do so, we analyzed a panel of 72 respiratory or blood isolates of P. aeruginosa collected from different patients in The First Affiliated Hospital, Zhejiang University School of Medicine Zhejiang, Hangzhou, which means that the exoU+ genotype mainly occurred in the east of China in this study.In total, 8 (11.1%) of 72 examined clinical isolates harbored exoU-like sequences by polymerase chain reaction (PCR)-based gene detection assay.The detailed information of the exoU+ phenotype is listed in Table 3.Among the exoU+ strains, five strains were blood and sputum isolates; the remaining three were isolated from bile, interstitial fluid, or urine, respectively.The cytotoxicity assay indicated that the cell viability of THP-1 macrophages infected by seven of eight exoU+ strains showed more than 40% cell toxicity compared to CT (uninfected cells) at 2 hpi, indicating the acute cytotoxicity of these strains.However, it is interesting to note that an isolate named PA AP showed 2.97% ± 0.33% for cell toxicity at 2 hpi, which was not significantly higher than CT (P > 0.05), indicating its low and non-acute toxicity against THP-1 cells (Fig. S3b).Thus, although the presence of exoU is associated with high virulence, it is not always the case.To further identify the basis for lack of high virulence in the strain PA AP, whole-genome sequencing (WGS) of these eight isolates was conducted.The results confirmed the presence of exoU in the eight strains.Notably, half of them (four of eight) belonged to ST463, which showed high virulence, and the strain PA AP belonged to ST1682.
In this study, the presence of exoU was not specifically associated with any source of infection, which is in accord with the observation in a previous study (39).The distribu tion of the exoU gene seems to vary significantly; for example, the ratio was 13% (4 of 32) (40) or 64% (9 of 14) (41) of bacteremia isolates, 10% (3 of 29) of CF isolates (42), and 61.5% (8 of 13) of microbial keratitis isolates, whereas none of the CF isolates (0/9) possessed this gene (43), 12% (43 of 189) of clinical isolates (39), 28% (32 of 115) of clinical and environmental isolates (44), and 52% (83 of 160) of clinical isolates (100% in wound) (45).The variations in the prevalence of the exoU+ strains among different studies could be ascribed to small numbers of samples analyzed or isolates not randomly selected or isolates from patients with different underlying clinical conditions with varied genotypes.

Mutation in spcU downstream of exoU may affect ExoU secretion and cytotoxicity in a clinical isolate
It is known that the increased virulence is associated with the secretion of ExoU, a toxin transported by the P. aeruginosa type III secretion system.However, not all strains of P. aeruginosa harboring the type III system (T3SS) genes are capable of secreting effector proteins (36).For example, 12 (34%) of the 35 examined P. aeruginosa isolates from patients with ventilator-associated pneumonia harbored the exoU gene, and immuno blot analyses showed that 83% of (10 of 12) strains produced detectable amounts of ExoU in vitro (35).Considering all the results together, we speculated that the strain PA AP showing low toxicity in this study could be due to inability to secrete ExoU in vitro, despite harboring the exoU gene.It has been shown that a small 15-kDa chaperone protein specific Pseudomonas chaperone for ExoU (SpcU), encoded by spcU downstream of exoU, forms a complex with ExoU and maintains the N-terminus of ExoU in an unfolded state, which is required for ExoU secretion (38,46).Indeed, our WGS data showed that the non-virulent strain PA AP contained the gene spcU, but there was a point mutation mutation of T→G at nucleotide position 280, causing phenylalanine (Phe) to be exchanged by valine (Val) at amino acid (aa) 94 of SpcU (SpcU-F94V) (Fig. 6), which may account for its loss of cytotoxicity in this strain.

Evolutionary characteristics of the exoU+ genotype
The confirmation of the role of exoU in virulence and prevalence prompted us to examine P. aeruginosa clinical isolates for its evolutionary characteristics.We selected 63 strains with known genome sequences combined with our sequenced seven strains to construct single-nucleotide polymorphism-typing phylogenetic tree analysis (Fig. 7a), to investigate the potential origin and the history of the spread of P. aeruginosa strains containing the exoU+ genotype.Phylogenetic tree analysis showed that clinical isolates from various sources do not cluster closely.MLST and serotype analysis revealed that the 70 isolates belonged to 49 types and 9 serotypes, demonstrating the sequence diversity of P. aeruginosa clones was high among different sources.The highly virulent strain BSI_S5 belonged to ST463/O4, which is consistent with a previous observation, where ST463/O4 isolate was found to be a potential high-risk clone (47).Remarkably, carbape nem resistance (CR; mainly blaKPC-2, blaOXA-396, and blaOXA-395) was identified in 51.43% (36 of 70) of the isolates and 63.64% (7 of 11) of ST463/O4.Among the ST463/O4, the other four strains would also be resistant to β-lactams because of the presence of intrinsic resistance genes blaOXA-486, along with the aminoglycoside-resistance gene (aph(3′)-IIb) and fluoroquinoloneresistance gene (crpP).Indeed, MIC determina tion confirmed that BSI_S5 was highly resistant to fluoroquinolone ciprofloxacin and β-lactams cefoperazone, aztreonam, piperacillin, ceftazidime, cefepime, piperacillin/tazo bactam, imipenem, and meropenem, but was susceptible to amikacin and tobramycin (Table S2).Thus, the ST463/O4 strain was coexisting with multidrug resistance genes, especially CR.

Mobile genetic elements in P. aeruginosa BSI_S5
Like many bacterial pathogens, P. aeruginosa genome is highly plastic and complex, composed of the core genome interspersed with DNA segments constituting mobile genetic elements (MGEs, a variety of accessory genes that form part of the pangenome).The MGEs embrace plasmids, bacteriophages, genomic islands, integrative and conjugative elements, integrative and mobilizable elements, insertion sequences (ISs), etc., which are considered as major contributors to the spread of antimicrobial resistance genes and virulence factors via horizontal gene transfer.To facilitate the understanding of specific regulatory mechanisms of exoU in BSI_S5, we analyzed MGEs present in its genome adjacent to exoU.As shown in Fig. 7b, there are two types of ISs present, IS222 and ISPa32, which belong to the IS3 family.Notably, exoU is located downstream of ISPa32.ISPa32 was originally detected in P. aeruginosa, with 80% (ORFA) and 95% (ORFB) amino acid similarity to IS407.Its complete sequences had three open reading frames (ORFs), and the third ORF is a putative ORFAB transposase (50).Further alignments showed that the IS sequences adjacent to exoU in BSI_S5 are quite similar to those of PA30, B14130, except some inversion or insertions of IS before the ISPa32 (Fig. 7b).Activation of neighboring gene expression can occur in two principal ways: either via promoters contained entirely within the IS driving transcripts that escape into neighboring DNA or by the formation of hybrid promoters following insertion (51).IS activity can result in increased resistance to antibiotics or insertional inactivation of specific porins (51,52).Besides, ISs are known to promote genome reshaping and may even affect the bacterial dependence on its host (51).ISs of P. aeruginosa, mainly those of the IS3 family, have been shown to be related to genome rearrangements (53).In the genome of BSI_S5, although two types of ISs (from the IS3 family) adjacent to exoU were observed, there is no clear evidence showing these ISs are implied in large-scale rearrangements.Nevertheless, the current analysis showed that ISPa32 is a species-spe cific sequence that may have the potential to transfer its neighboring gene (i.e., exoU) to other strains via horizontal gene transfer.Further exploration of the relationship between ISPa32 expression and transposition activity is essential to understanding the dynamics of ISs in exoU expression and transfer in P. aeruginosa.

DISCUSSION
In this study, we used the comparative genomic analysis followed by gene knockout and complementation to investigate the unique genes affecting the pathogenicity of P. aeruginosa.We found that exoU, chr_1696, and chr_4238, which encode the T3SS effectorExoU, T1SS permease/ATPase or polysaccharide biosynthesis protein, respec tively, are specifically present in the highly virulent P. aeruginosa BSI_S5 strain, which belongs to ST463/O4.Deletion of exoU showed significantly attenuated cytotoxicity, while deletion of the two other unique genes caused no apparent loss of cytotoxicity compared to the parent strain BSI_S5; and complementation of ΔexoU mutant restored its cytotoxicity.
To verify the role of ExoU during infection, P. aeruginosa BSI_S5 (WT) and ΔexoU were employed in a bloodstream infection utilizing a mouse model via tail vein.The increased survival rate of mice infected with ΔexoU compared to WT showed that ΔexoU was significantly attenuated for virulence in vivo, and the complementation of the ΔexoU mutant completely restored its virulence.Bacterial recovery showed that a homogeneous distribution of the P. aeruginosa strain in the liver when inoculated with low dose of parent strain or ΔexoU, which indicated that deletion of exoU gene in the strain did not affect its liver survival but only changed its survival in the lung and spleen.No P. aeruginosa spp.were detected in the GBs and feces whether infected by low dose of parent strain or ΔexoU.With high dose of P. aeruginosa, there was a significant difference in the bacterial loads between the two groups (spleen and feces), and the bacterial population in the GBs was dramatically expanded in the majority of mice infected by parent strain while not in the ΔexoU-infected mice.Regardless of low or high dose used, there was also a significant difference in bacterial population in the lung or spleen, while the bacterial numbers in GBs or feces differed significantly between parent strain and ΔexoU-infected mice.It could be speculated that during bacteremia, a small subpopulation of P. aeruginosa replicated in the liver and seeded the GB.Once in the GB, this population then expanded dramatically, disseminated to the intestines, and was excreted in the feces.Following the liver-GB-intestinal excretion pathway, P. aeruginosa from systemic infections could return to the environment in high numbers in order to enhance transmission between mice (34).Moreover, the hematoxylin and eosin (H&E) staining showed that ExoU induced severe lung lesions and the mucosal damage to GB.Taken together, these results demonstrated that ΔexoU was attenuated in vivo compared with the parent P. aeruginosa strain, underscoring it as a strong candidate for further targeting to improve treatment of P. aeruginosa systemic infections.This supports that T3SS is a major driver of virulence in P. aeruginosa and is known to damage epithelial surfaces through direct injection of cytotoxic effector proteins.Moreover, ExoU can significantly contribute to the colonization of P. aeruginosa in the lung, liver, spleen, GB, and feces of the host with acute bloodstream infection.
We also found that a strain PA AP harbored the exoU gene but showed low cytotoxic ity.This indicates the presence of exoU gene by PCR detection alone is not sufficient for identifying ExoU expression strains for virulence testing.Instead, an intact protein SpcU encoded by spcU downstream of the exoU is crucial for ExoU secretion and virulence expression, as we identified a clinical isolate with intact exoU gene but had a point mutation T→G at nucleotide position 280 (mutation from Phe to Val at amino acid 94) in spcU that caused loss of cytotoxicity.Since strains harboring and expressing exoU may cause particularly severe disease in patients, it would be of interest to develop a rapid molecular test to detect the high virulence phenotype by detecting not only the presence of the exoU gene but also intact downstream spcU by targeted sequencing of both genes.It is worth noting that although SpcU was previously found to be essential for ExoU secretion by sequential deletion of the gene loci using lab experimental system (38), this study is the first to identify a point mutation in SpcU in a clinical isolate responsible for the loss of its function in causing defective cytotoxicity.Although our results clearly show that the intact coexistence of spcU and exoU are required for exerting toxicity of P. aeruginosa, one limitation in this study is that we could not elucidate how the amino acid change affects the functional domains in SpcU that interrupts the secretion of ExoU.Further experimental work will be necessary to elucidate the nature of the ExoU and SpcU interaction through structural biology studies.
In addition, our results indicate that clinical isolates from various sources do not cluster closely and belong to different ST types.Notably, the highly virulent strain BSI_S5 is serotype ST463/O4, which is an emerging high-risk clone that spread rapidly in East China, whose genome commonly contains exoU and multidrug resistance genes, especially CR.The exoU gene is present in all of the ST463/O4 genomes but not specific for this lineage.The combined extensive drug resistance and the high virulence due to exoU may contribute to its high pathogenicity and mortality in the host and spread in hospital environments.Besides, the two types of ISs (IS222 and ISPa32 belonging to the IS3 family) adjacent to exoU may potentially influence exoU transfer to different strains.Finally, the exact mechanism by which ExoU induces cellular death and its role in pathogenesis of P. aeruginosa ST463/O4 will be further investigated in the future.
In conclusion, through a series of whole-genome sequencing, molecular biology and cellular and animal studies, we found that the type III secretion system effectorExoU was the main determinant of pathogenicity of the highly virulent P. aeruginosa clinical isolate BSI_S5, which caused serious systemic infections and belonged to serotype ST463/O4.We showed that deletion of exoU but not two other unique genes, chr_1696 and chr_4238, caused significant attenuation of cytotoxicity and virulence, while complemen tation of the ΔexoU mutant restored cytotoxicity.Interestingly, we found that a clinical isolate with intact exoU gene but harboring a mutation of T→G substitution at nucleotide position 280, causing amino acid Phe to Val change at aa 94 of SpcU (SpcU-F94V), showed significant loss of cytotoxicity, suggesting that a functional downstream SpcU is required for ExoU secretion and cytotoxicity.In addition, we assessed a total of 77 (5 + 72) clinical isolates of P. aeruginosa and found that exoU+ genotype comprised 11.68% (9 of 77) of these strains (excluding 63 strains with known genome sequences and PAO1).Our study demonstrates the important role of not only ExoU but also its downstream intact SpcU in conferring virulence of P. aeruginosa and calls for surveillance of both exoU and spcU by targeted sequencing for improved detection and control of virulent P. aeruginosa infections in the clinical setting.

Bacterial strains
Human sputum, blood, or stool samples from patients diagnosed with bloodstream or lung infections were obtained from an already existing collection of The First Affiliated Hospital, Zhejiang University School of Medicine, in 2020.The identity of P. aeruginosa was confirmed by Pseudomonas isolation agar (PIA) plates and 16S ribosomal DNA sequencing.

Whole-genome sequencing
Whole-genome sequencing of clinical isolates of P. aeruginosa and the common model strain P. aeruginosa PAO1 included as a control strain was conducted by Personalbio (Jiangsu, China).Briefly, the WGS was performed using Illumina NovaSeq sequencing platform and also PacBio Sequel sequencing platform in order to obtain accurate and complete whole genome sequences.A total of two libraries were constructed for each strain for the WGS sequencing.

Measurement of virulence using chicken embryo infection model
Chick embryo infection model was used to evaluate the virulence of P. aeruginosa strains.Hatching eggs were obtained from a hennery in Deqing, Zhejiang Province.A batch of 10-day-old chick embryos were used in this study.A hole (1.2-mm diameter) was aseptically drilled in the egg shell using a needle, and a suspension (100 µL, 10 5 CFU/ egg) of P. aeruginosa strains being tested was inoculated intra-allantoically, followed by aseptic sealing of the inoculation hole with sterile medical pressure-sensitive adhesive tape.Two controls were included, including eggs inoculated with PBS (100 µL) and eggs not subject to any inoculation treatment.Each run contained 10 eggs for each bacterial strain and each control as described (21).The eggs were checked by ovoscope and candle at different time intervals to evaluate the viability of the embryos.Dead embryos were identified by the absence of spontaneous movement (usually associated with hemorrhage), loss of vascular architecture, and abnormal morphology.Such eggs were removed from the incubator, and the stage of development was assessed using Hamburger and Hamilton's key.Incubation was terminated on the 17th day of embryo genesis, when all live embryos were euthanized.

Cytotoxicity assay
The cytotoxicity assay was performed as described with some modifications (54).Briefly, six-well tissue culture plates were seeded with 5 × 10 5 THP-1 cells/mL cultured in RPMI 1640 medium (supplemented with 10% fetal bovine serum (FBS), 100-U/mL penicillin and streptomycin).Cells were maintained at a density of 1 × 10 5 -10 6 /mL and differenti ated by addition of 100-ng/mL phorbol 12-myristate 13-acetate 24-48 hours prior to infection.The cells were washed two times with PBS to remove non-adherent cells and fresh RPMI 1640 medium (supplemented with 10% FBS) was added (for lactate dehydro genase assay, fresh RPMI 1640 medium was supplemented with 1% heat-inactivated FBS).Cells were incubated at 37°C under humidified atmosphere and 5% CO 2 at least 30 min prior to infection with P. aeruginosa.THP-1 cell cytotoxicity was determined by incubation of cells with overnight cultures of P. aeruginosa at a multiplicity of infection (MOI) of 10:1 in 2.0 mL of assay medium.The infected cells were incubated for 3.5 hours at 37°C and 5% CO 2 before cell viability (trypan blue stain) was evaluated.Supernatants were harvested and analyzed for LDH activity using an LDH Activity Assay Kit (Beyotime Biotechnology, Shanghai, China) according to the manufacturer's instructions.

Flow cytometry analysis
THP-1 cells (5 × 10 5 cells/mL) were infected with P. aeruginosa for different times (MOI = 10:1).Cells in medium alone were set as the control.Apoptotic cells were investigated by double staining with annexin V/PI using an annexin V-EGFP apoptosis detection kit (Cellor Lab, Shanghai Epizyme Biomedical Technology Co., Ltd) according to the manufacturer's instructions.Cells were centrifuged at 200 × g for 3 min, and washed twice with ice-cold PBS (pH 7.4).The cell pellet was resuspended in 200 µL of bind ing buffer.Cells were stained by adding 4 µL each of FITC-annexin V and PI working solutions.After incubation at room temperature for 10 min in the dark, samples were analyzed by flow cytometry (CytoFLEX, Beckman) within 1 hour.
Partially overlapping flanking primers were designed to amplify the upstream and the downstream 5,000-bp DNA sequences of chr_1696, exoU, and chr_4238 of P. aeruginosa (Table S1).The upstream and downstream of the gene of interest were amplified by PCR as follows: 10 ng of upstream primer and downstream primer, 100 ng of template DNA, 25 µL of 2× Phanta Flash Master Mix (Vazyme, Nanjing, China), added to ddH 2 O to a final volume of 50 µL.The thermo cycling parameters were programmed according to the following protocol: 98°C for 30 s, then 30 cycles at 98°C for 10 s, Tm for 5 s, and 72°C for 15 s, then a final extension at 72°C for 1 min.pK18mob-sacB plasmid was propagated in E. coli using LB medium containing 50-µg/mL kanamycin and incubated at 37°C for 16 hours.The cultures were collected and the plasmid DNA was purified using TIANprep Mini Plasmid Kit (TIANGEN, Bei jing, China) according to the manufacturer's instructions.To generate the linearized overlapping fragments of pK18mob-sacB, DNA was digested with SaII (Thermo Fisher).The PCR products and the linearized vector were analyzed by DNA agarose gel electro phoresis and purified by gel extraction kit (TIANGEN) and dissolved in TE buffer, then quantified by Nanodrop (Thermo Fisher).The upstream and downstream of each gene were inserted into pK18mob-sacB using ClonExpress II One Step Cloning Kit (Vazyme) to obtain pK∆1696, pK∆exoU, and pK∆4238, respectively.The newly assembled products were transformed into E. coli DH5α competent cells (Vazyme) via heat shock.Subse quently, triparental matings were used to mobilize plasmids from E. coli DH5α to P. aeruginosa with the conjugative helper strain E. coli HB101 (pRK2013) (55).The cells were plated on PIA plates containing 2,000-μg/mL kanamycin (PIA/Kan 2,000), and incubated at 37°C overnight.Colonies growing on the PIA plates were checked by colony PCR using the pK18mob-sacB forward primer and a primer specific to the genomic DNA adjacent to the cloned 500-bp downstream sequence (Table S1).Positive colonies were purified by streaking on new (PIA/Kan 2,000) plates.Purified merodiploids were streaked on TYS10 plates and incubated at room temperature (~20°C) for 72-96 hours.Colonies were examined by colony PCR with primers specific to the genomic DNA before the cloned 500-bp upstream sequence and after the cloned 500-bp downstream sequence of each deleted locus (Fig. S1).Colonies with the confirmed deletion were purified by streaking on a TYS10 plate and further verified by Sanger sequencing (Tsingke Biotech Co., Ltd.; Fig. S4-1 through S4-3).

Gene complementation
For gene complementation, the locus of exoU as well as its upstream and downstream was amplified using appropriate primers (Table S4), purified, and seamlessly fused with the linearized pK18mob-sacB.The resulting plasmid was transformed into E. coli DH5α.Colonies containing the complement construct were confirmed by colony PCR using the universal M13 primer pair.Positive colonies were further selected for DNA preparation and sequencing.The fused plasmid was then transformed into the ΔexoU mutant by electroporation followed by two rounds of screening (the same procedure as above).Finally, two populations were identified on the TYS10 plates: one included the desired complementation strain [ΔexoU (exoU)], while the other included wild type revertants (ΔexoU).As a result, the complementary strain ΔexoU (exoU) carries a 2,064-bp DNA fragment exoU, and the sequences were verified by DNA sequencing (Fig. S4-4).

FIG 2
FIG 2 Circular representation and gene family analysis of the P. aeruginosa genome.Circular representation of the P. aeruginosa BSI_S5 genome.From inside out, the innermost circle indicates the chromosomal location in base pairs; the second circle represents the GC sketch; the third circle represents the G + C content; circles 4 and 7 show the classification of CDS into functional categories based on clusters of orthologous groups of proteins; circles 5 and 6 show the location of CDS, tRNAs, or rRNAs on each chromosome in blue (a).Gene family analysis of PA14, PAO1, S1, COPD_S2 (S2), BSI_S3 (S3), AECOPD_S4 (S4), and BSI_S5 (S5).Petal map of gene family of different strains of P. aeruginosa (b) and proportion distribution map of pan-genome and dispensable genome of P. aeruginosa (c).

FIG 3
FIG 3 Location of virulence-associated genes in BSI_S5.A total of 442 virulence factors are identified in BSI_S5, and they belong to 13 different categories, which are marked in different colors.

FIG 5
FIG 5 ExoU is critical for virulence and pathogenicity of P. aeruginosa infection.The Kaplan-Meier survival curve for C57BL/6 (7-to 9-week-old) female mice challenged with the P. aeruginosa exoU deletion mutant (Δexou, n = 11), ΔexoU complementation strain [δexou (exoU), n = 14], or parent P. aeruginosa BSI_S5 (WT, n = 14) strain.Comparison of the survival curves was performed using log-rank (Mantel-Cox) test (a).Female BALB/c mice (n = 7) were intravenously injected with ~5 × 10 6 CFU of P. aeruginosa BSI_S5, Δexou (b-d).Female BALB/c mice (n = 6) were intravenously injected with ~4 × 10 7 CFU of WT, δexou (e-j).Mice were euthanized at 22 hpi (n = 6), and bacteria were enumerated from several anatomical sites by plating serial dilutions of organ homogenates (CFU per gram of organ weight) and the recovered bacteria are presented (star on the x-axis means the number of CFU was below the detection limit (<10 CFU/g).Organs where no bacteria were recovered are denoted with a diamond on the x-axis.Lung (b and e), spleen (c and f), liver (d and g), blood (h), gallbladder (i), and feces (j).For input 1 g equals 1 mL of wet weight.GB and lung embedded in formalin were evaluated by hematoxylin and eosin staining (k and l).WT, wild type.

FIG 6
FIG 6 Demonstration of a point mutation in the spcU gene.The wild-type spcU gene sequence in P. aeruginosa BSI_S5 and the other seven exoU+ clinical isolates with defined virulence (a).A point mutation of T→G at nucleotide position 280 in the spcU gene from the isolate named PA AP, causing phenylalanine to be replaced by valine at aa 94 of SpcU (SpcU-F94V) (b).

FIG 7
FIG 7 The ST463/O4 clone harboring the exoU and multidrug resistance genes and the genome organization of exoU loci with insertion sequence IS3 families.Full-genome single-nucleotide polymorphism-based phylogenetic and characteristics of all isolates.Strains sequenced in this study are marked with asterisk (a).IS families adjacent to exoU in BSI_S5 genome as defined in Prokka (b).

TABLE 1
Features of P. aeruginosa genomes a ORF, open reading frame.b "/" represents no plasmid detected.

TABLE 2
Unique virulence-related genes identified in strain BSI_S5 exoU contributes to P.

TABLE 3
The MLST genotype of exoU+ P. aeruginosa clinical isolates from patients a a MDS, myelodysplastic syndrome.