Bacterial strains and plasmids used in this study were listed in Table S1 . PU-1 is an O2:K1 ExPEC strain (isolated from the blood of a piglet) causing acute sepsis in mouse infection model (Ma et al., 2020; Ma et al., 2021). All strains were grown on Luria-Bertani (LB) broth medium at 37°C with 180 rpm, supplemented with corresponding antibiotics, or isopropyl–D-thiogalactopyranoside (IPTG) when necessary. DNA amplification, ligation and electroporation were performed as previously described (Ma et al., 2018) unless otherwise indicated. Deletion mutants were constructed using the λ red mutagenesis method (Datsenko and Wanner, 2000), and the details of primers, restriction enzymes and fragments’ deletion have been listed in Table S2 . All restriction and DNA-modifying enzymes were purchased from Thermo Fisher Scientific (Waltham, MA, USA) and performed according to the supplier instruction.

Porcine blood and mucus of the respiratory tract were collected from the healthy pigs of a slaughterhouse to perform the following studies. Blood from healthy donators was obtained from Jiangsu province Blood Center, and related experiments were approved by the Medicine Human Subjects committee of Jiangsu Province. Five-week-old female specific pathogen free (SPF) BALB/c mice were purchased from Yangzhou University (Comparative Medicine Center). All animal experiments were performed in strict accordance with the animal welfare standards of the Animal Research Committee Guidelines of Jiangsu Province (License Number: SYXK (SU) 2017-0007), and approved by the Ethics Committee for Animal Experimentation of Nanjing Agricultural University.

Ten mice in each group were challenged by intraperitoneal injection with the indicated strain at the designed doses and monitored for symptoms until 7 days post-infection. The negative-control group was challenged with an equal volume of sterile PBS. To evaluate bacterial proliferation in vivo, the bacterial load assay was conducted. Five mice in each group were inoculated with 1 × 10⁶ CFU/mouse of the indicated strain, the infected blood was harvested at the designed time of post-infection, and then serially diluted in PBS and plated on LB agar to enumerate the CFU.

Total RNA was extracted with the E.Z.N.A. bacteria RNA isolation kit (Omega), and residual genomic DNA was then removed by digestion with DNase I (TaKaRa). The PrimeScript RT reagent kit (TaKaRa) was used for cDNA synthesis. The RT-qPCR was performed using SYBR premix Ex Taq (TaKaRa) with the gene-specific primers. The relative amount of target gene mRNA was normalized to the transcript of housekeeping gene tus (Ma et al., 2018), and the relative fold change was calculated by the threshold cycle (2^-ΔΔCT) method. The reported values represented the mean ± SD of three independent RNA extractions.

The recombinant His₆-Vat^PU-1 and His₆-Tsh^PU-1 proteins were purified by Ni-NTA Spin Columns (QIAGEN) from BL21 (DE3) carrying the recombinant pET-21a plasmid after IPTG induction. Human hemoglobin (Sigma-Aldrich) prepared at a concentration of 50 μg/mL in PBS were coated onto separate wells of a 96-well plate. The purified His₆-Vat^PU-1 and His₆-Tsh^PU-1 were added to each well for 1 h at 37°C, and then washed three times and incubated with the anti-His antibody (Abcam, 1:2000) at 37°C for 2 h. After thrice wash, the processed membranes were stained with the HRP conjugated secondary antibodies (Thermo Fisher, 1:2000) at 37°C for 1 h, and detected using the 3,3’-diaminobenzidine. The reaction was stopped after 30 min by addition of 1 M H₂SO₄, and absorbance was measured at 450 nm. Three independent experiments were performed, with four wells for every reaction in each experiment, and the values obtained were averaged.

The mucin gel penetration assay were used as previously described (Henderson et al., 1999; Silva et al., 2003). A solution containing 10% mucin of porcine respiratory tract and 0.3% agarose in HBSS was loaded into a 1-mL injection syringe, creating a mucous column. A 0.1 mL prepared bacterial cells (10 × 10⁹ cells/mL) were layered onto the mucin. The columns were incubated for 3 h at 37°C in a vertical position. Afterwards, five fractions (each one contains 0.2 mL) were collected from the button by applying gentle pressure. Each fraction was serially diluted and plated to the LB agar media for CFU enumerating. Western blot assays were performed to analyses the cleavage activity of SPATEs to mucin. The mucins were extracted from the mucus of porcine respiratory tract as previously described (Gibold et al., 2016). Degradation reactions were separated by 12% SDS-PAGE, and then transferred to PVDF membranes (Bio-Rad) for subsequent blocking, washing, incubating with the specific and secondary antibodies (using as the instructions of manufacturer), and detecting using the 3,3’-diaminobenzidine.

Before incubating with the conjugated mAb of APC-CD43 (Invitrogen), PMNs were incubated with the human IgG to block Fc receptors. The prepared cells were incubated with SPATEs for the indicated times, and then stained by incubation with the dye-conjugated antibodies specific to the extracellular domain of host glycoproteins. The positive staining with antiCD16 and CD16b mAbs (R&D Systems) were selected as the low forward-scatter and high side-scatter characteristics, respectively, and then used to gate the neutrophils. The samples were analyzed in an Accuri C6 fow cytometer (BD Accuri)/fluorescence-activated cell sorter (FACS), and analyzed using the CFlow plus software (BD Accuri).

Chemotaxis and transmigration assays were performed according to previously described protocols (Yamamoto et al., 2000; Ruiz-Perez et al., 2011). For the chemotaxis assay, 3 × 10⁵ calcein-stained human PMNs were incubated with the indicated protein in the upper chamber of Transwell, and 100 mM of IL-8 (MedChemExpress) or 100 nM of fMLP (N-formyl-methionineleucine-phenylalanine, MedChemExpress) was added to the lower chamber as chemoattractant. After 4 h incubation, the PMNs that had migrated toward the chemoattractant were collected for counting with a Fluoroskan fluorometer. For transendothelial migration assays, human brain microvascular endothelial cells (HBMECs) were seeded on the inserts of Transwells at a concentration of 2 × 10⁴ cells/well, and cultured until the monolayers were confluent (~ 4 days). Afterwards, similar operating steps were performed as the chemotaxis assays.

The blood samples were analyzed by a Blood RT (Routine Test) machine in the Animal Hospital of Nanjing Agriculture University. The whole bloods of mice infected with the indicated bacterial strains were collected, and centrifuged to get the sera. The mouse interleukin 6 and 8 ELISA test kits (HUYU biological technology Co Ltd, shanghai) were used as the instructions of manufacturer to detect the levels of interleukin 6 and 8 release in the sera.

Statistical analyses were performed using Prism 8.0 (GraphPad), and the full details were described below. Two-way ANOVA was used for the qRT-PCR assay, One-way ANOVA was performed for the results of bacterial survival assay, blood routine test, IL-6 and IL-8 release, PMNs chemotaxis and transmigration. For infection experiments, survival data were analyzed with the log rank test. For all tests, a P value < 0.05 were considered statistically significant, and all data were shown as mean ± SD.

PU-1 is an O2:K1 ExPEC strain causing acute sepsis in mouse infection model (Ma et al., 2020; Ma et al., 2021). The curve of blood bacterial loads in mice infected by strain PU-1 showed the maximum value (more than 10⁸ CFU/mL) at 12 h post-infection ( Figure 1A ), which was remarkably higher than the infection groups of virulent ExPEC strain DCE7 from phylogenetic group D (Zhu et al., 2017), and strain DCE1 from group A. Furthermore, the blood of mice (with obvious clinical symptoms: tremble, orbital hemorrhage, anorexia, ataxia, anaesthesia et al.) in PU-1 infection group showed a significantly slighter decrease in bacterial loads after 12 h post-infection than the mice challenged with strains DCE7 and DCE1 ( Figure 1A ). The blood routine testing results showed that white blood cell (WBC) and neutrophil cell (NEU) counts of peripheral blood from the PU-1 infection group were significantly less than that of DCE7 infection group ( Figures 1B, C ) at the 6 h post-infection (no significant differences in blood bacterial loads between PU-1 and DCE7 infection groups at this point in time). These data suggested that strain PU-1 has the potential to modulate the host immune responses for its optimal blood infection in vivo.

Comparative analysis based on our previous transcriptome data (Ma et al., 2020) found two special genes, FQU83_21385 (chromosome encoding) and FQU83_01245 (plasmid encoding), which encoded two serine protease autotransporters of enterobacteria (SPATEs), and showed the significant upregulation of more than 3 folds in animal blood but not in serum ( Figure 2A ). Phylogenetic analysis showed that the FQU83_21385 and FQU83_01245 belong to the class-2 of SPATEs ( Figure 2B ), and share highest sequence identities with the well-studied virulence factors Vat homologue of strain LF82 and Tsh homologue of strain APEC O1, respectively, thereby were redesignated as Vat^PU-1 and Tsh^PU-1 in strain PU-1. To explore whether these two SPATEs involving in ExPEC’s bloodstream survival, non-polar deletion mutant strains were constructed under the background of wild-type strain PU-1. The results of bacterial counting demonstrated that the inactivation of Vat^PU-1 alone or Tsh^PU-1 alone did not affect the bacterial survival in host blood ( Figure 2C ). It should be noted that mutant strain with double deletions of vat^PU-1 and tsh^PU-1 significantly decreased survival in host blood, while had no significant effect for the bacterial survival in host serum ( Figure 2C ). These results suggested that at least one SPATE in strain PU-1 was required to resist the clearance mediated by immunocytes within blood, while was not involved in resistance to the bactericidal effects of serum.

An animal infection test using the BALB/c mice was employed to further certify the pathogenic roles of these two SPATEs during bloodstream infection. As shown in Figure 3A , the mice infected by the Δvat&tsh^PU-1 , but not the single deletion mutants Δvat^PU-1 and Δtsh^PU-1 , showed a significantly higher survival rate (80%), compared with the 100% death of mice infected by the wild-type strain. Consistently, only the double deletions of vat^PU-1 and tsh^PU-1 significantly attenuated the bacterial loads in mice blood compared with the wild-type strain ( Figure 3B ). Porcine and human bloods were then used to assess the survival rates of indicated ExPEC strains. The results demonstrated that the mutant strain with double deletions of vat^PU-1 and tsh^PU-1 significantly decreased survival in both of these two types of blood ( Figures 3C, D ), while the inactivation of Vat^PU-1 alone or Tsh^PU-1 alone did not affect the bacterial survival here. These data indicated that Vat^PU-1 and Tsh^PU-1 are critical for the full virulence and blood infection in ExPEC strain PU-1.

Numerous class-2 SPATEs of pathogenic E. coli have been reported to bind to and cleave mucin-like proteins for optimal colonization (Henderson et al., 1999; Dutta et al., 2002; Leyton et al., 2003; Harrington et al., 2009), including Pic, Tsh, Vat, et al. In this study, both of Vat^PU-1 and Tsh^PU-1 were identified to have the abilities to bind to hemoglobin ( Figure 4A ), a well-known mucin-like glycoprotein widely presenting in red blood cells (RBCs). We then investigated the ability of Vat^PU-1 and Tsh^PU-1 to degrade mucus by using a column penetration assay. Unlike Δvat&tsh^PU-1 , wild-type strain PU-1 penetrated through the entire mucus column (fractions 1 to 5, Figure 4B ). In the fractions 3 and 4 (middle of the column), the efficiency of Δvat&tsh^PU-1 was about two orders of magnitude lower than that of PU-1 strain at penetrating the mucus column. In order to investigate the mucinolytic activity of Vat^PU-1 and Tsh^PU-1, we examined the ability of related strains to hydrolyze porcine respiratory mucins by Western blotting. The cultural supernatant of wild-type strain PU-1 was able to degrade respiratory mucins unlike those of Δvat&tsh^PU-1 ( Figure 4C ). Overall, these results suggest that Vat^PU-1 and Tsh^PU-1 promote the bacterial penetration through the mucus layer by altering the gel-forming mucins.

SPATEs have been confirmed to target a broad range of mucin-like glycoproteins present not only on the RBCs and epithelial cells, but also on the surface of leukocytes. To determine whether Vat^PU-1 and Tsh^PU-1 cleaved such proteins, the CD43, a major cell surface glycoprotein expressed on leukocytes, was addressed. We performed flow cytometry analyses of human PMNs treated with purified Vat^PU-1, Tsh^PU-1 and denatured proteins to further confirm the CD43 cleavage on cell surface. Staining with anti-CD43 mAbs revealed that the extracellular CD43 was present on the surfaces of PMNs after treatment with denatured Vat^PU-1 and denatured Tsh^PU-1 ( Figure 4D ), while was cleaved, becoming undetectable on cells treated with purified Vat^PU-1 and Tsh^PU-1. These results, coupled with the previous study verifying the cleavages of CD44, CD45, CD93 and CD162 by the Tsh homologue from Avian Pathogenic E. coli (APEC) (Ruiz-Perez et al., 2011), suggested that Vat^PU-1 and Tsh^PU-1 have the potential to cleave diverse O-glycosylated mucin-like proteins located on peripheral blood leukocytes, which may facilitate the bacterial survival within host bloodstream.

We next explored whether the cleavage of mucin-like glycoproteins mediated by Vat^PU-1 and Tsh^PU-1 disturb the function of peripheral blood leukocytes during ExPEC infection. Here chemotaxis and transmigration assays of PMNs were performed according to previously described protocols (Ruiz-Perez et al., 2011; Ma et al., 2018). For the chemotaxis assays, IL-8 or fMLP within the lower chamber of Transwell were used to stimulate the translocation of PMNs through an abiological membrane from the upper chamber, and PMNs that migrated in the lower chamber were measured. The movements of PMNs stimulated by IL-8 and fMLP were significantly enhanced compared with the untreated PMNs in this model ( Figure 5A ), while incubation with the Vat^PU-1 but not the denatured Vat^PU-1 was significantly reduced the activated movement to the similar levels with the untreated PMNs. Similar results were also observed when the PMNs were incubated with the Tsh^PU-1 ( Figure 5B ). These observations suggested that both two SPATEs can impair the PMNs’ chemotaxis. For the transmigration assays, the PMNs within the upper chamber of Transwell were stimulated to transmigrate through the endothelial cell monolayers, and PMNs in the lower chamber were measured after 4 h incubation. Expectedly, the transmigration of PMNs incubated with Vat^PU-1 or Tsh^PU-1 was significantly reduced compared with the buffer alone as negative control ( Figures 5C, D ); denaturation of Vat^PU-1 or Tsh^PU-1 by heating completely abolished the above effects. Taken together, these observations suggest that the PMNs’ functions were significantly impaired by the interaction with Vat^PU-1 and Tsh^PU-1 during the bloodstream infection of ExPEC.

Further study managed to investigate the potential modulations of SPATEs to the immune responses in a mouse infection model. Blood routine testing results showed that white blood cell (WBC) and neutrophil cell (NEU) counts of peripheral blood from Δvat&tsh^PU-1 infection group were significantly higher than that of wild-type infection group ( Figures 6A, B ). The cytokines’ detection was then performed. Compared with wild-type strain, the inactivation of Vat^PU-1 alone or Tsh^PU-1 alone did not affect the release levels of IL-6 and IL-8 in blood ( Figures 6C, D ). However, the double deletion mutant Δvat&tsh^PU-1 induced the significantly higher levels of IL-6 and IL-8 at the 6 h post-infection (no significant differences in blood bacterial loads compared with the wild-type at this point in time), indicating that a greater inflammatory response was activated compared with the wild-type infection. These data suggested that Vat^PU-1 and Tsh^PU-1 coordinately attenuate the host immune responses, thus may mediate the evasion of ExPEC strain PU-1 from the immune clearance of blood leukocytes.