HvKP expresses aerobactin in the murine gut but it is dispensable for gut colonization

Previous studies have demonstrated that of the four siderophores expressed by hvKP strains, aerobactin is the most abundant and a critical virulence determinant [17,21]. The aerobactin operon is well-conserved in hvKP isolates and encodes four biosynthetic genes (iucA-D) and the outer membrane receptor (iutA) [25,26] (Fig 1A). The hvKP1 clinical isolate (ST86, KL2) harbors the ~ 200 Kb virulence plasmid pVir/pLVPK, which includes the aerobactin (iuc1) gene cluster [27]. Since siderophores play a key role in aiding bacteria to acquiring iron in iron-scarce environments [17,22,28], we examined the growth kinetics of an aerobactin biosynthesis mutant, iucA::kan (iucA-) in the hvKP1 background, under both iron-replete (LB) and iron-depleted (LB with 200 µM 2, 2’-Dipyridyl [DIP]) conditions. Compared to the wild-type (WT), the iucA- mutant had a modest but consistent growth defect in iron-depleted media, whereas no growth defect was observed in iron-replete media (S1A Fig). In competitive co-culture experiments, neither strain outcompeted the other under either growth condition (S1B Fig), suggesting trans-complementation of the iucA- mutant by WT, likely due to the presence of an intact iutA gene encoding the aerobactin outer membrane receptor. Our in vitro qRT-PCR results, comparing RNA isolated from iron-replete and iron-depleted medium, demonstrated that aerobactin gene expression is highly upregulated in the WT strain, more so than the other siderophores (S2A Fig). Loss of aerobactin did not result in an increase in expression of other siderophores except for salmochelin (iroB) expression displaying a significant 4-fold increase, relative to the WT. Nevertheless, total siderophore output was significantly lower in the iucA- mutant, as measured using chrome azurol S (CAS) agar (S2B Fig).

Download:

• PNG
  larger image
• TIFF
  original image

Fig 1. Aerobactin is dispensable for gastrointestinal colonization but required for efficient translocation and high virulence.

(A) Genetic organization of the iuc operon, encoding aerobactin synthesis and receptor genes, in the hypervirulent K. pneumoniae clinical isolate hvKP1. The iucA gene is shown in blue and iutA (receptor) in purple. (B) Schematic of the murine oral feeding model of Kpn gastrointestinal colonization and the experimental timeline. (C) RNA was isolated from cecal contents of C57BL6/J mice (n = 4) 4-6 days post-inoculation (PI). The log₁₀ fold-change in transcripts was measured via qRT-PCR for iucA (aerobactin), entC (enterobactin), iroB (salmochelin), and irp2 (yersiniabactin) and plotted to compare hvKP1 siderophore gene expression in the murine gut (in vivo) to in LB media (in vitro). Expression of gyrA and rpoD served as controls, and all data was normalized to the K. pneumoniae 16S housekeeping gene for 2^-ΔΔC_T analysis. Each biological replicate (n = 4) was run in duplicate, and their averages with SEM is shown. Kruskal-Wallis test, followed by Dunn’s test of multiple comparisons compared to gyrA control, was performed. (D-E) C57BL6/J mice were orally inoculated with either the WT (gray) or the isogenic iucA- mutant (blue) and followed for 7 days (n ≥ 10 per group). (D) Fecal shedding was plotted with each symbol representing a single mouse on the indicated day and the bars specifying the median bacterial shedding. The dashed line represents the limit of detection (LOD). (E) Kaplan-Meier curves compare mice survival between the WT (grey) and the isogenic iucA- mutant (blue) inoculated mice (n ≥ 10 per group). Log-rank (Mantel-Cox) tests were performed to compare survival odds. LOD, limit of detection. (F) Translocation kinetics of mice orally fed 10⁶ CFU of the WT strain. Shown is the bacterial burden in the spleen (S), liver (L), kidneys (K), and cecum (Ce) at 12 hours (n = 5), 18 hours (n = 12), 24 hours (n = 8), and 30 hours (n = 11) PI. The percentage of mice with bacteria detected in the extra-intestinal organs (S, L, K), suggesting translocation, is listed at the top of the graph for each time point. The LODs for the samples are depicted as dotted lines, and the pertaining organs are indicated. (G) Bacterial burden in the extra-intestinal organs of mice at 24 hours PI with either the WT (n = 10) or the iucA- mutant (n = 8). (D, G) Mann-Whitney U tests were used to determine statistical significance. LOD, limit of detection. *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001; ns, not significant.

https://doi.org/10.1371/journal.ppat.1014122.g001

Given the role of the GI tract as a reservoir for Kpn [6,7,29], and a transposon mutagenesis screen with an intact microbiome implicating siderophores as important gut colonization determinants [30], we investigated whether hvKP1 expresses siderophores during gut colonization. Using our oral-feeding mouse model of Kpn gastrointestinal colonization [29] (Fig 1B), we harvested the cecal contents of mice post-inoculation (PI) with hvKP1 and isolated bacterial RNA. Compared to RNA isolated from broth-grown cultures, qRT-PCR of cecal RNA samples revealed elevated levels of all four siderophores, with aerobactin and enterobactin showing significant upregulation (Fig 1C). These findings establish that there is likely competition for iron in the gut causing upregulation of aerobactin by hvKP residing in the murine gut.

Next, as siderophores have been demonstrated to be important for gut colonization for enteric bacteria during inflammation and antibiotic treatment [31–33], and as aerobactin is expressed by hvKP1 in the mouse GI tract, we tested the iucA- mutant in our mouse model and determined the ability of this mutant strain to colonize the murine GI tract. As fecal shedding correlates with density of Kpn in the gastrointestinal tract [29], shedding was used as a non-invasive metric of determining gut colonization. Over the course of the 7-day study, both the WT and the iucA- mutant shed at equivalent levels (Fig 1D), indicating that aerobactin synthesis is not required for colonization. Interestingly, compared to the WT, mice inoculated with the iucA- mutant displayed reduced mortality, displaying a 70% increase in survival (Fig 1E).

Thus, our findings implicate aerobactin in promoting growth under iron-limited conditions and show that, while it is actively expressed in the gut, it is dispensable for GI colonization but contributes to hvKP virulence.

Aerobactin contributes to translocation of hvKP1 from the gut

Previously, we observed that mice that succumbed to infection following GI inoculation with hvKP1 harbored high bacterial burdens in the liver, spleen, and kidneys [29]. This suggested that bacterial translocation had occurred. To better understand the translocation process, we monitored hvKP1 translocation kinetics by quantifying bacterial loads in the gut and peripheral organs (spleen, liver, kidneys) from 12 to 30 hours PI. At 12 hours PI, hvKP1 was primarily confined to the GI tract. By 18 hours PI, bacteria were detectable in extra-intestinal organs in approximately half the mice, and by 24 hours PI, hvKP1 was recovered from the spleen, liver, and kidneys of all colonized mice (Fig 1F). Given that the iucA- mutant colonized the gut similarly to WT (Fig 1D) but exhibited reduced virulence (Fig 1E), we hypothesized that its attenuation was likely due to an impaired ability to translocate from the gut to extra-intestinal sites. Consequently, we selected 24 hours PI to compare bacterial burden at sterile sites between the WT and the iucA- mutant. At 24 hours PI, we observed reduced burden of the iucA- mutant compared to the WT isolate, in the spleen, liver, and kidneys (Fig 1G). These results identify a potential role for aerobactin in the translocation process.

As the WT strain complemented the growth defect of the iucA- mutant under iron-poor growth conditions (S1B Fig), we envisaged the WT strain compensating for the in vivo translocation defect of the iucA- mutant. We inoculated mice at a 1:1 ratio of the WT and the mutant strain, while both strains colonized the GI tract, the WT maintained a modest competitive advantage. Specifically, the competitive index (CI) in the cecum and colon leaned toward negative values, with mean offsets of -0.14 and -0.19, respectively, suggesting the mutant was slightly outcompeted even in the gut. However, this advantage was more pronounced in the extra-intestinal sites, where the WT strain outcompeted the aerobactin mutant in 4 out of 5 mice, rather than complementing it (Fig 2A). We speculate that, compared to the gut, the differences observed in the systemic sites could be due to the infection barriers present within the host. Additionally, since bacterial burdens varied among WT-colonized mice in both the gut and extra-intestinal sites, we considered whether a correlation existed between the bacterial loads in the gastrointestinal tract and those in the extra-intestinal organs post-translocation. At 24 hours PI, we identified a strong positive correlation between the liver and the cecum, followed by the spleen and the cecum (Fig 2B), suggesting gut bacterial burden impacts translocation. These results are in line with a recent Kpn molecular study that established that bacterial burden in the lung can impact metastatic spread to secondary sites [34].

Download:

• PNG
  larger image
• TIFF
  original image

Fig 2. hvKP dissemination from the gut to systemic sites requires high bacterial burden and does not rely on the lymphatic system.

(A) Competition assay in which mice (n = 6) were orally fed a 1:1 mixture of the WT (gray-filled shapes) and the iucA- mutant (blue open shapes); each mouse is represented by a unique symbol. Bacterial burden was enumerated from the gastrointestinal samples (F: fecal, CE: cecum, CO: colon) and the extra-intestinal organs (S: spleen, K: kidneys, L: liver) 24 hours post infection (PI). A dotted line connects bacterial densities recovered from the same animal for the specified sample. Competitive index (CI) values were calculated as the log₁₀ average CI for each sample. The dotted lines depict the LOD for the respective samples. (B) Correlation analyses of extra-intestinal (spleen, liver, and kidneys) bacterial burden of hvKP1 relative to gastrointestinal (cecum) colonization at 24 hours PI, displayed as a correlation matrix using the Pearson correlation coefficients (r) (i) and as simple linear regressions (ii). R squared for each regression is listed. (C) CFU obtained from the mesenteric lymph nodes (MLN) of mice orally inoculated with the WT and enumerated at 12 hours (n = 5), 18 hours (n = 12), 24 hours (n = 8), or 30 hours (n = 11) PI. A dotted line marks the LOD. (D) Bacterial densities in the MLN at 24 hours PI in mice orally fed either the WT (gray circles) or the iucA- mutant (blue squares). The dotted line indicates the LOD. (E) Colonization density of the iucA- strain in the (i) gastrointestinal tract (ileum, cecum, and colon) and (ii) the extra-intestinal sites (blood, spleen, and liver) of mice at 4 days, 15 days, and 21 days PI. Statistical significances of CIs were calculated using Wilcoxon signed-rank test with a theoretical median of 0. PI, post-inoculation. *P < 0.05, ns, not significant.

https://doi.org/10.1371/journal.ppat.1014122.g002

Since several enteric pathogens, including Salmonella enterica serovar Typhimurium, Shigella spp., and Listeria monocytogenes, exploit M cells for systemic dissemination [3], we examined the mesenteric lymph nodes (MLNs) in hvKP1-colonized mice. In contrast to reports with S. Typhimurium [35], hvKP1 was rarely detected in the MLNs, despite being present in the extra-intestinal organs (Figs 1F, 1G, 2C-2D), suggesting a vascular route for translocation. Furthermore, as mice colonized with iucA- still exhibited translocation, with the majority of mice surviving to day 7 (Fig 1E), we wondered whether the surviving mice remained colonized in extra-intestinal sites or cleared iucA- over time. We followed mice for either 4, 15, or 21 days PI and observed that the iucA- mutant stably colonized the gut through to the study endpoint at 21 days (Fig 2Ei). In contrast, colonization of the extra-intestinal sites declined over time, and by day 21, all mice had cleared the bacteria from systemic sites (Fig 2Eii).

Next, we evaluated the significance of aerobactin in context with the other siderophores produced by hvKP1. We utilized an isogenic strain that only produces aerobactin, entB::cam irp2::hyg, and a strain that lacks all four siderophores, ΔiucA entB::cam irp2::kan. The aerobactin-only strain colonized the murine gut similarly to WT, whereas the siderophore-null strain had a defect in colonization (Fig 3A), indicating that siderophores contribute to robust gut colonization. Interestingly, the siderophore-deficient mutant was completely avirulent in our murine model (Fig 3B), whereas the aerobactin-only mutant was partially virulent when compared to the WT strain, implying the importance of all four siderophores to hvKP virulence. Collectively, our results demonstrate that hvKP can efficiently translocate from the GI tract to extra-intestinal sites, causing invasive disease, with aerobactin contributing to this process and influencing virulence.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 3. Aerobactin is sufficient for gut colonization but other siderophores are required for increased virulence.

(A-B) C57BL6/J mice were orally inoculated with either the WT, an isogenic aerobactin-only synthesizing mutant (entB::cam irp2::hyg), or a siderophore-deficient mutant (ΔiucA entB::cam irp2::kan) and followed for up to 7 days. (A) Median fecal shedding for the WT and each of the isogenic mutants on the designated day, with each symbol representing a single mouse on the given day (n = 9 per group). The dotted line represents the LOD. A Kruskal-Wallis test, followed by Dunn’s test of multiple comparisons was performed for statistical significance. (B) Kaplan-Meier curve indicating the percent survival of the mice. Log-rank (Mantel-Cox) tests were performed to compare survival odds. *P < 0.05, **P < 0.01, and ***P < 0.001; ns, not significant.

https://doi.org/10.1371/journal.ppat.1014122.g003

Having observed a difference in bacterial burden at systemic sites post-GI colonization (Fig 1G), we speculated that this difference could either be due to a defect in translocation, or poor survival of the iucA- strain in the extra-intestinal space. A recent transposon mutagenesis screen determined Kpn bacterial factors important for bacteremia and identified several TonB-dependent iron uptake systems, but none of the siderophore systems were required for colonization of the spleen after systemic infection [36]. Thus, we also used a systemic infection mouse model, in which 10⁶ CFU were injected directly into the bloodstream via the lateral tail vein, bypassing the need for translocation (Fig 4A). No discernable differences in bacterial burden were observed in the systemic sites (spleen, blood, liver, and kidney) between the WT and iucA- strains at either 6 or 24 hours PI (Fig 4B, 4C). Furthermore, similar to the survival kinetics post-gut colonization (Fig 1E), we observed reduced mortality in mice orally inoculated with the iucA- mutant compared to the WT strain (Fig 4D), providing further evidence that aerobactin impacts hvKP virulence. To further probe whether the iucA- mutant had a survival or growth defect in the systemic sites, we evaluated how well it survived in the presence of 10% rabbit serum. Unlike a capsule-deficient mutant, KPPR1S ΔwcaJ [37], iucA- grew similar to WT in rabbit serum (Fig 4E). These results likely preclude early clearance of the iucA- as the reason for reduced burden in the extra-intestinal sites, and instead raise the possibility that the reduced bacterial burden at systemic sites is due to a translocation defect.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 4. Aerobactin is dispensable during early colonization events but contributes to virulence in a systemic model of infection.

Eight-week-old C57BL6/J mice were injected with 10⁶ CFU in 100 µL of either the WT or the isogenic iucA- mutant via the lateral tail vein. (A) Schematic of the murine systemic infection model. (B-C) Bacterial burden in the extra-intestinal sites (spleen, blood, liver, and kidneys) was either examined at 6 hours (B) or 24 hours (C) PI (n = 5-7 per group). Symbols with a red outline represent mice at an in-extremis state (severe illness) at 24h PI. Groups were compared using the Mann-Whitney U tests. (D) Survival analysis of mice infected via lateral tail vein with either WT (dashed gray line) or the iucA- mutant (solid blue line). Survival was monitored for 8 days, and statistical significance determined using a Log-rank (Mantel-Cox) test with P-values indicated. ns, not significant. (E) WT (gray), iucA- (blue) or KPPR1S ΔwcaJ (capsule-less control [magenta]) grown to mid-log phase (10⁷ CFU) in LB + 200 µM DIP (iron-chelated media) were exposed to 10% rabbit serum, 10% heat inactivated (HI) rabbit serum, or PBS-control, for 1 hour at 37°C to determine survival. Shown is the mean ± SEM (n = 4) each assay was performed in triplicate and statistical significance was determined using the Kruskal-Wallis test, followed by Dunn’s test of multiple comparisons. LOD, limit of detection; PI, post-inoculation. *P < 0.05, ns, not significant.

https://doi.org/10.1371/journal.ppat.1014122.g004

Aerobactin-mediated virulence is conserved across genetically diverse hvKP isolates

Having established that aerobactin is critical for translocation and virulence in hvKP1, we next investigated whether this function is conserved across genetically diverse hvKP strains. Kpn exhibits substantial genetic heterogeneity, with over 130 distinct capsule types circulating across classical and hypervirulent pathotypes [38,39]. The hvKP1 isolate is a K2 capsule type, and since the most prevalent capsule types for hvKP strains are K1 and K2 [40], we examined additional clinical isolates, including hvKP2 (ST23, KL1) and hvKP94 (ST23, KL1), along with their isogenic aerobactin-deficient counterparts (iucA::kan). We observed no difference in fecal shedding from mice colonized with hvKP2 or hvKP2 iucA- (Fig 5A). The hvKP94 isolate is considered as a partially virulent strain [23], and hvKP94 iucA- also behaved similarly in the gut (Fig 5B). In accordance with other murine challenge studies [23], compared to hvKP2, the partially virulent strain hvKP94, in our murine model of infection had reduced virulence. Conversely, both mutant strains presented reduced virulence compared to their respective wild-type parental counterparts (Fig 5C, 5D). Consistent with our observations with hvKP1 (S2B Fig), disrupting iucA in both hvKP2 and hvKP94 backgrounds significantly reduced total siderophore output (Fig 5E). These results reinforce that while aerobactin is dispensable for gut colonization, it is a conserved driver of translocation and virulence across genetically distinct hvKP lineages.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 5. Aerobactin impacts virulence without affecting hvKP gut levels across genetically distinct isolates.

(A-B) C57BL6/J mice were orally inoculated with either the hvKP2 (KL1 isolate) or the isogenic iucA- mutant (A) or hvKP94 (KL1 isolate) and its isogenic iucA- mutant (B), and fecal shedding was used to measure gastrointestinal bacterial burden. All strains were grown overnight in LB, and mice inoculated with ~10⁶ CFU. Median shedding is shown for respective groups on the given day (n = 10 per group). A dotted line indicates the LOD. Statistical analysis was done using the Mann-Whitney U tests. (C-D) Survival analyses of mice colonized in the gastrointestinal tract with hvKP2 (C) or hvKP94 (D) (dashed lines) and their iucA- isogenic mutant counterparts (solid line) are shown. Survival of mice was monitored for 15 days PI and an in-extremis state or death was scored as non-survival. Statistical significance was determined using the Log-rank (Mantel-Cox) tests with P-values indicated. (E) Total siderophore production was assessed on CAS agar plates for either hvKP2 or hvKP94 and the respective iucA- isogenic mutants. Each strain was grown in LB overnight and diluted 1:10 in PBS before plating. Halos were measured and expressed as the mean area in pixels x1000. Shown are the averages for each experiment (n = 5). Statistical significance was determined using the Mann-Whitney U tests. LOD, limit of detection. *P < 0.05;, **P < 0.01, ns, not significant.

https://doi.org/10.1371/journal.ppat.1014122.g005

Aerobactin promotes hvKP adherence, invasion, and translocation

Since our results precluded a major role for the M cells in hvKP translocation (Fig 2C, 2D), and previous studies have established that hvKP isolates interact with colonic epithelial cells [13,41], we sought to understand the impact of aerobactin on hvKP interactions with intestinal epithelial cells using a colonic Caco-2 cell culture model. We did not observe a difference in the ability to adhere or invade between the WT and the iucA- mutant when grown in iron-replete media (LB) (S3A, B Fig). In contrast, when both strains were grown in conditions where aerobactin provides a growth advantage (LB-DIP), we observed a significant reduction in both adhesion and subsequent invasion of the epithelial cells by the iucA- mutant in comparison to the WT strain (Fig 6A, 6B). When adjusted for adhesion, the defect in invasion was masked (S3C Fig), indicating that the observed invasion defect is primarily due to a defect in adhesion, which is considered a critical first step in translocation. The same effect in adhesion and invasion was observed in the hvKP2 and hvKP94 backgrounds (S3 D, E Fig). Furthermore, this defect in adhesion which impacts invasion was rescued upon trans-complementation of the iucA- mutant with plasmid pFUS2[iucA-D] containing the iuc operon, which partially restored siderophore production [17] (S3 F-H Fig). Subsequently, we used Caco-2 cells grown on transwell inserts to assess whether the reduction in adhesion and invasion affected bacterial translocation through a monolayer, by measuring CFU that have crossed through the epithelial barrier. Compared to the WT, the iucA- mutant strain had a defect in translocation (Fig 6C), suggesting that the initial adhesion defect likely results in reduced translocation-ability of the iucA- strain. Trans-complementation with pFUS2[iucA-D] once again rescued the defect in translocation (S4A-D Fig).

Download:

• PNG
  larger image
• TIFF
  original image

Fig 6. Aerobactin promotes hvKP1 adherence, invasion, and translocation.

(A-B) Caco-2 (human intestinal epithelial cells) monolayers were infected with either the WT (gray) or the iucA- mutant (blue) grown to mid-log phase in LB + 200 µM DIP (iron-chelated media) to test for adhesion (A) or invasion (B) at an MOI 50. Data are expressed as percent to WT, and the results of 3 independent assays, each with 3-5 replicates, are shown as mean ± SEM. Statistical significance was determined using the Mann-Whitney U tests. (C-D) Polarized Caco-2 cells were grown on transwell inserts and allowed to differentiate for 19 days before infecting with either WT (gray) or iucA- (blue) grown to mid-log phase in LB + 200 µM DIP. The top and bottom chambers were sampled at 30 min and 2h PI to measure for translocation (C), with data expressed as percent to WT. The results of 4 independent assays, each with ≥ 5 replicates, is shown and Mann-Whitney U tests were performed for statistical significance. (D) TEER measurements were obtained using a Millicell ERS-2-Volt-ohmmeter at the indicated time points and plotted as mean percent of baseline (starting value). Wells treated with EGTA or DMEM served as experimental controls. The pooled data from 5 independent assays are shown ± SEM, each with ≥4 replicates. (E). Representative immunohistochemistry (IHC) images (i) of murine proximal colons (5 µm sections) that were collected 24h PI from mice inoculated with WT or iucA- mutant or from uninfected vehicle-control mice (n ≥ 2 per group) and stained for occludin (1:400), and quantified H-scores (0-300) (ii) detailing occludin expression via Halo image analysis. PI, post-inoculation. *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001.

https://doi.org/10.1371/journal.ppat.1014122.g006

Moreover, our transwell assay allowed us to provide insight into hvKP translocation route. Two well-established routes of translocation are known: transcellular (moving through cells) and paracellular (moving between cells). To assess which route was being utilized, we measured tight junction integrity using transepithelial electrical resistance (TEER). Interestingly, no differences in TEER values between the WT and iucA- mutant strains were identified (Fig 6D), suggesting that translocation likely initiates through the transcellular route. To confirm the effects of hvKP on tight junction proteins in vivo, we examined the expression of Occludin, a key protein in tight junction formation and function [42,43], in murine colonic tissue. The expression of Occludin in both WT-infected and iucA- mutant-infected mice was similar to the expression in naïve (control) mice (Fig 6E), indicating that the tight junction proteins were unperturbed by hvKP during early infection (24 hours PI), with or without aerobactin. Histological examination of the hematoxylin and eosin (H&E)-stained colonic tissues showed only minor changes in mice infected with the WT or the iucA- mutant compared to the naïve mice, as indicated by both image analysis and blinded scoring (S4E-Fii Fig). Taken together, our data illustrate the role of aerobactin primarily in adherence to epithelial cells, a prerequisite step for translocation through the transcellular route.

Aerobactin deficiency affects the HMV phenotype in hvKP

Bacterial surface features play a crucial role in engagement with the host environment [44,45]. Fimbrial structures on the bacterial surface promote adhesion and, in turn, facilitate host cell invasion [46]. Kpn expresses two major types of fimbriae: type 1 and type 3. The type 1 fimbriae mediate adhesion to host epithelial cells in the respiratory and urinary tracts, whereas the type 3 fimbriae are linked to biofilm formation and adherence to abiotic surfaces under static conditions and both type 1 and 3 have been implicated in catheter associated bladder infection [47]. To assess whether aerobactin influences fimbrial expression under iron-depleted conditions, we measured the expression of fimI and fimA, which encode a putative anchor subunit and the major shaft protein of the type 1 fimbriae, respectively. We also analyzed the expression of mrkA and mrkB, which encode the major pilin subunit and a chaperone protein important for type 3 fimbrial assembly. Transcriptional analysis via qRT-PCR revealed that although all four genes were upregulated under low-iron conditions, deletion of iucA did not affect their expression (S5A Fig). These results suggest that the reduced epithelial cell engagement observed in the iucA- mutant is likely independent of fimbrial gene expression.

Another surface feature that can impact adherence to epithelial cells is capsular polysaccharide (CPS). Capsule has been shown to interfere with adherence in multiple pathogens [48–50]. Thus, we explored whether aerobactin contributes to CPS and the related hypermucoviscous (HMV) phenotype of hvKP. HMV, a hallmark of hvKP isolates, is characterized by increased mucoidy due to longer CPS chain-length distribution [10] and is regulated by the rmpD gene [51] within the rmp locus (Fig 7A). To test whether iron levels impact expression of the CPS and the rmp loci, we conducted qRT-PCR analysis using RNA isolated from the WT and the iucA- mutant grown in either iron-replete or iron-depleted media. Genes involved in capsule biosynthesis and retention (galF, wzi, and manC) and the rmp locus (rmpA, rmpD, and rmpC) were upregulated in iron-depleted conditions, with an even more pronounced increase observed in the iucA- mutant (Fig 7B, 7C). Sedimentation assays, revealed significantly increased HMV levels in iucA- mutants compared to their WT counterparts (Figs 7D, S5B), while CPS levels, measured by uronic acid quantification, were unaffected (S5C Fig). Furthermore, trans-complementation of the iuc operon restored HMV levels to those observed in the WT strain (S5D Fig), implicating aerobactin in modulating hvKP HMV phenotype.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 7. Aerobactin expression under iron restrictive conditions regulates HMV levels.

(A) Schematic of the rmp operon of K. pneumoniae, with RmpA positively regulating the rmp locus, RmpD modulating HMV and RmpC positively regulating capsule biogenesis (B-C) qRT-PCR comparing gene expression of capsule locus genes (galF, wzi, and manC) with rpoD transcript levels shown as a control (B) and rmp locus genes (rmpA, rmpD, and rmpC) (C) in the WT and the isogenic iucA- mutant grown in LB + 200 µM DIP (iron-chelated media) and LB (iron-rich media). Each biological replicate (n = 5) was run as a duplicate. The fold-change in transcripts is shown with gyrA as the internal control for 2^-ΔΔC_T. Statistical significance was determined using the Mann-Whitney U test. (D) HMV comparison via low-speed spin-down assay of the WT and iucA- mutant grown to mid-log phase (OD₆₀₀ ~ 0.5) in LB + 200 µM DIP (iron-chelated media), expressed as the percent to WT. Shown are the results from 3 independent experiments (n = 3). (E-F) Caco-2 monolayers were infected with either the WT (gray), iucA- mutant (blue), rmpD::kan (dark gray), or the ΔiucA, rmpD::kan (dark blue) that were grown to mid-log phase in iron-chelated media and compared for adhesion (E) and invasion (F) at an MOI of 50, and expressed as a percent to WT ± SEM. (D-F) The results from 3-4 independent assays are depicted (n ≥ 2). Mann-Whitney U tests were performed on the pooled samples of each experiment to determine statistical significance. *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001; ns, not significant.

https://doi.org/10.1371/journal.ppat.1014122.g007

Based on our HMV results, we focused on RmpD, as it is known to directly interact with Wzc, a tyrosine kinase involved in capsule polymerization and export, leading to longer CPS chains characteristic of HMV [10,52]. Previous work showed that an rmpD mutant deficient in HMV exhibited increased adherence to J774A.1 macrophage cells [51]. Therefore, we hypothesized that if the reduced adherence observed in the iucA- mutant is due to increased HMV mediated by RmpD, then a double mutant lacking both iucA and rmpD should display similar adherence levels similar to an rmpD- mutant. Compared to the WT, we observed an increased level of adhesion and corresponding invasion in our cell culture model with an rmpD::kan (rmpD-) mutant (Fig 7E, 7F), consistent with previous data using J774A.1 macrophage cells [51]. Furthermore, the ΔiucA rmpD- mutant behaved similarly to the rmpD- mutant (Fig 7E, 7F), suggesting that dysregulation of the rmp locus in the absence of aerobactin contributes to reduced adherence and invasion of the mutant. Taken together, our gene expression and cell culture data demonstrate that aerobactin-mediated iron acquisition plays an important role in modulating capsular polysaccharide and HMV gene expression, thereby influencing hvKP interactions with epithelial cells.

Ethics statement

This study was conducted in accordance with the guidelines outlined by the National Science Foundation’s animal welfare requirements and the Public Health Service Policy on Humane Care and Use of Laboratory Animals [83]. All murine studies were conducted according to the American Association for Laboratory Animal Science (AALAS) guidelines and were approved by the Wake Forest Baptist Medical Center Institutional Animal Care and Use Committee (IACUC). The approved protocol numbers for this project are A20-084 and A23-064.

Bacterial strains

Strains, plasmids, and primers used in this study are listed in S1–S3 Tables, respectively. The in-frame deletion of iucA in hvKP1 (WT) (ST86, KL2; bacteremia, hepatic and splenic abscess; Buffalo, New York, USA) background was generated as described previously [84]. Briefly, the kanamycin (kan) cassette with flanking FLP recombinase target sites (FRT) was amplified from pKD4 with 50 bp homology to iucA at the 5’ and 3’ ends, using Q5 polymerase (M0491L; New England Biolabs), and treated with DpnI (1 µL/ 100 µL PCR product) to digest the pKD4 template. The purified PCR product was electroporated (1.8 kV, 400 Ω, 25 µF) into hvKP1 containing the temperature-sensitive pKD46 plasmid, which encodes the λ red recombinase genes under the control of an arabinose-inducible promoter (pBAD). Successful mutants were selected on lysogeny broth—Lennox (LB) agar plates supplemented with kan (25 µg/mL). The presence of the kan cassette was confirmed via PCR. The kan cassette was removed with Flp recombinase using pFlp3 [85] to generate the ΔiucA strain. The plasmid complemented strain was generated by transforming hvKP1 iucA::kan with the pFUS2[iucA-D] plasmid [17]. Empty pFUS2 vectors were transformed into both WT and hvKP1 iucA::kan for direct comparison.

To construct the siderophore-deficient strain, ΔiucA entB::cam irp2::kan, in the hvKP1 background, an irp2::kan mutant was first generated in the ΔiucA background, following the same λ red recombinase process listed above, but with a 60 bp homology for the irp2 gene at the 5’ and 3’ ends. Subsequently, the entB::cam cassette was amplified using Q5 polymerase (M0491L; NEB) from isolated genomic DNA from an entB::cam mutant in MKP103 [86]. The purified PCR product was electroporated into ΔiucA irp2::kan background strain containing pKD46. Successful mutants were selected on LB agar supplemented with chloramphenicol (cam) (50 µg/mL), confirmed via PCR, and plasmid cured at 37°C.

Generation of the rmpD mutants in the WT and ΔiucA backgrounds involved amplification of the rmpD::kan cassette with flanking FRT sites from pKD4 with 60 bp homology to rmpD at the 5’ and 3’ ends, using Q5 polymerase (M0491L; NEB), and treated with DpnI (1 µL/ 100 µL PCR product) to digest the pKD4 template. As described above, the purified PCR product was used with WT and ΔiucA strains containing pKD46 for λ red recombination. The resulting mutants were selected on LB agar supplemented with kan (25 µg/mL) and confirmed via PCR.

The hvKP2 and hvKP94 iucA::kan mutants (AZ235 and AZ255) were generated using λ red recombinase via pKD46 to disrupt the iucA gene with the kan cassette from pKD4, as described above. Recombinants were selected on LB agar supplemented with kan (25ug/mL) and confirmed via PCR. The sequence type (ST) and capsule type (KL) for the parental hvKP2 (ST23, KL1) and hvKP94 (ST23, KL1) isolates were derived from the original genomic analysis conducted by the source laboratory (Russo Lab, Buffalo, New York, USA).

The apramycin-resistant hvKP1, hvKP1 attTn7::apra (AZ314), was constructed by amplifying the apramycin resistance cassette from the attTn7 site of AZ94 [87] and introducing it into hvKP1 (AZ121) via λ red recombination, as described above. Recombinants were selected on LB agar containing apramycin (apra) (50 µg/mL) and confirmed by PCR.

Unless otherwise specified, all strains were grown overnight in LB at 37°C with constant agitation. Overnight cultures were diluted to 10⁷ CFU/mL for mouse infections or subcultured 1:100 for other assays.

Kinetic growth assays

Strains to be tested were grown overnight at 37°C with constant agitation in LB-Lennox, OD₆₀₀ adjusted to 3, and then subcultured 1:100 in either LB or LB with 200 µM 2,2’-dipyridyl (DIP) to chelate iron. Subcultures were grown at 37°C with constant agitation, and at the designated time points, a sample was removed, diluted, and plated for enumeration.

For in vitro competitive growth studies, overnight cultures were prepared as above. Competing strains with distinct antibiotic selection markers were diluted 1:100 into fresh test media, and grown at 37°C with constant agitation. Samples were collected at designated time points, and plated on selective antibiotic plates for enumeration of each strain. The Competitive Index (CI) was calculated for each time point using the following formula, where t(0) is the CFU/mL at the time of initial subculture and t(x) is the CFU/mL at different times post-initial subculture.

Mucoviscosity assay

Mucoviscosity levels were measured using the low-speed spin-down assay, previously described [88], with slight modifications. Strains to be tested were grown overnight in LB as described above, subcultured 1:100 in fresh LB with or without 200 µM DIP, and further grown at 37°C with constant agitation until mid-log (OD₆₀₀ ~ 0.5). Samples were then adjusted to OD₆₀₀ of 0.5 in a final volume of 1.2 mL, and OD₆₀₀ measurements were recorded in triplicate. Afterward, 1 mL of each sample was centrifuged (1,000 x g for 5 min), and OD₆₀₀ measurements were recorded again in triplicate. The ratio of the average of the three measurements post-centrifugation over the average of the three measurements before centrifugation was quantified. Strains transformed with pFUS2 plasmids were grown in media supplemented with gentamicin to prevent plasmid loss.

Uronic acid quantification assay

Uronic acid (UA) content, as an indicator of capsular polysaccharide amount, was measured as previously described [88,89] with slight modifications. Overnight cultures were subcultured 1:100 in the indicated media and grown to mid-log at 37°C with constant agitation. UA was extracted from each 500 µL sample by adding 100 µL of 1% Zwittergent 3–08 in 1M citric acid and incubating for 20 min at 50°C. Afterward, samples were centrifuged (21,000 x g for 5 min) at room temperature, and 300 µL of supernatant was incubated overnight with 1.2 mL of 100% ethanol (200 proof). Precipitates were centrifuged (21,000 x g for 10 min) at 4°C, the supernatants were removed, pellets were air-dried and subsequently resuspended in 200 µL of H₂O, and incubated for 10 min at 100°C with 1.2mL of 12.5mM sodium tetraborate in sulfuric acid. Samples were then chilled on ice for 10 min and incubated with 20 µL 0.15% 3-phenylphenol. Each sample was loaded as triplicate on a 96-well plate, and absorbance was measured at 520 nm on a plate reader (Biotek H1 Synergy). UA levels were quantified against a standard curve with glucuronic acid.

Mouse infections and sample collection

C57BL/6J specific pathogen-free (SPF) mice were obtained from Jackson Laboratory (Bar Harbor, ME) and bred and maintained at Biotech Place, Wake Forest Baptist Medical Center animal facility. For all studies, mice were randomly allocated to different groups, and the sample size was not predetermined.

Oral inoculations were performed as previously described [29]. Briefly, 5–7 week-old mice were fasted for 3–4 hours and orally fed 100 µL of Kpn resuspended to ~10⁶ CFU/100 µL in 2% sucrose-PBS via pipette in two doses of 50 µL an hour apart. The inoculum dose was serially diluted and plated on selective media to enumerate and confirm the dose. For survival studies, animals were monitored for 7 days, and a state of in extremis or death resulted in the study endpoint.

For systemic infections, 8-week-old C57BL/6J SPF mice were injected intravenously with ~10⁶ CFU of Kpn in PBS by tail vein inoculation. Following inoculations, animal health and body weight were monitored throughout the study period. Humane endpoints were defined by clinical signs of moribundity, including altered gait, hunched posture, increased respiration, or a failure to explore the cage when disturbed. While weight loss was monitored as an indicator of decline, it was assessed in the context of the animal’s overall clinical stability; mice that remained active and responsive despite transient weight loss were closely monitored. Mice requiring euthanasia were marked as a death in the survival curve. All mice that survived up to day 8 post-infection were euthanized using CO₂ (2 L/min for 5 min) followed by cardiac puncture.

Bacterial shedding in the feces was quantified as previously detailed [29,30,89–91]. At designated time points, the bacterial burden in the gastrointestinal tract (distal ileum, cecum, and proximal colon) and extra-intestinal organs (mesenteric lymph nodes (MLN), liver, spleen, and kidneys) was determined as described. Briefly, samples were collected from mice following euthanasia, weighed, and diluted with 1X PBS. For the liver and kidneys, equal weight-to-volume of PBS was added, and all other organs were diluted 1:10. Diluted samples were homogenized using the Fisherbrand Bead Mill 24 Homogenizer at a strength of 3.1 for 2 min. Blood was collected from the submandibular vein using a 5 mm lancet. Samples were serially diluted and plated on antibiotic plates to enumerate CFU counts.

In vivo competition studies were conducted as previously described [29,30,89–91], with a 1:1 ratio of each strain resuspended in 10⁶ CFU/100 µL in 2% sucrose and orally fed to the mice as described above. Collected and homogenized samples were serially diluted and plated on selective antibiotic plates to distinguish the two strains. The CI was calculated using the following formula, where the output is the CFU/mL quantified in the sample and the input is the CFU/mL of the infection dose.

RNA extraction, cDNA synthesis, and RT-qPCR

RNA was isolated from Kpn strains from in vitro samples using the TRIzol method [92] with modifications [89]. Briefly, overnight cultures grown in LB were subcultured 1:100 into their respective media and grown to OD₆₀₀ ~ 0.5 before RNA isolation.

Bacterial RNA was isolated from murine cecal samples as previously described [30,89,91,93]. Briefly, cecal contents from mice colonized with the hvKP1 isolate (n = 4) were collected 4–6 days post-infection and placed in a 2 mL screw cap tube containing glass beads, diluted 1:1 with RNAlater (Invitrogen, AM7020), homogenized, and stored overnight at 4°C. An equal volume of chilled 1X PBS was added to the samples, and they were centrifuged (700 x g for 1 min at 4°C), and the supernatants were collected and centrifuged (900 x g for 5 min at 4°C). The pellets were resuspended in 500 µL 2X Buffer A (200 mM NaCl, 200 mM Tris base, 200 mM EDTA), 210 µL 20% SDS, and 500 µL UltraPure Phenol:Chloroform:Isoamyl Alcohol (25:24:1). Isolated crude RNA from both in vitro and in vivo samples were treated with DNase (Invitrogen; AM1907) followed by cDNA synthesis (iScript, BIO-RAD) [89,91,92]. Samples were purified via the Qiagen MinElute PCR Purification Kit.

RT-qPCR was performed as previously detailed [94,95] using Kpn-specific primers [89]. DNA gyrase (in vitro) and 16S rRNA (in vivo) served as reference genes with primers specific to Kpn. Samples were run in duplicate on a CFX384 Touch real-time PCR detection system (BIO-RAD). RNA expression was quantified using the ΔΔC_T threshold cycle (CT) method with fold-change calculated using 2^-ΔΔC_T [89,94]. For in vitro samples, 3–5 biological replicates were tested in duplicate. For each in vivo sample, the average of two independent experiments was plotted.

Adhesion and invasion assays

Adhesion and invasion assays were carried out as previously described [13], with slight modifications. Approximately 5 x 10⁵ Caco-2 [Caco2] (ATCC HTB-37) colonic epithelial cells in Dulbecco’s Modified Eagle Medium (DMEM) supplemented with 10% heat-inactivated Cytiva HyClone fetal bovine serum (FBS), 1% Cytiva HyClone Non Essential Amino Acids (NEAA) 100X solution, 1% GlutaMAX supplement, and without antibiotics, were seeded in 1 mL volume per well in 24-well plates (Corning Costar Flat Bottom Cell Culture Microplates), and allowed to adhere and form monolayers overnight at 37°C with 5% CO₂. Cells were washed with 1 mL of 1X Dulbecco’s Phosphate Buffered Saline (DPBS), and 1 mL fresh media without antibiotics was added to each well. Kpn was grown to mid-log in either LB or LB supplemented with 200 µM DIP, OD₆₀₀ adjusted to 0.5, and resuspended in FBS-free DMEM. Strains transformed with pFUS2 plasmids were grown in media supplemented with gentamicin. Bacteria were added to each well in a multiplicity of infection (MOI) of 50 bacteria/cell in 250 µL. Plates were centrifuged at 200 x g for 5 min. Samples of each inoculum were serially diluted and plated on antibiotic plates for CFU counts.

For adhesion assays, the centrifuged plates were incubated for 20 min at 37°C with 5% CO₂. Afterwards, the wells were washed three times with 1X DPBS, and the bacteria were recovered with 500 µL of 0.2% Triton X-100 (Sigma, X100-100ML). For invasion assays, plates were incubated as described above for 2 hours and subsequently washed twice with 1X DPBS. Cells were then treated with 1 mL of media containing 100 µg/mL gentamicin and incubated for another 2 hours to kill any extracellular bacteria. Cells were washed twice afterwards with 1X DPBS and intracellular bacteria were recovered with 500 µL of 0.2% Triton X-100.

Recovered bacteria were serially diluted and enumerated on selective media plates. The number of adhered or invaded bacteria was calculated as the proportion of the inoculum. Three to four independent experiments were conducted, each with ≥ 2 replicates. The data was expressed as a percent to WT.

Translocation assays and TEER measurements

Translocation assays were performed as previously described [13,96–98] with modifications. Caco-2 cells (~4x10⁵ in 0.5 mL) in complete media (DMEM supplemented with 10% heat-inactivated FBS, 1% Gibco Antibiotic-Antimycotic (100X), 1% NEAA, 1% GlutaMAX supplement, and 50 µM 3-mercapto-1,2-propanediol) were seeded onto transwell inserts (3 µm pore size, 1.12 cm² surface area) in 12-well plates (Corning Transwell Multiple Well Plate with Permeable Polycarbonate Membrane Inserts) and allowed to differentiate for 19 days. Kpn strains were grown and resuspended in FBS-free DMEM at half the original volume to concentrate each inoculum. Cells were washed twice with media without antibiotics. Bottom chambers were subsequently filled with 1.5 mL of media without antibiotics, and top chambers (apical side of differentiated cells) were inoculated with 500 µL of bacteria (1 x 10⁸ CFU). Plates were incubated for 30 min as described above. Samples were collected from the top and bottom chambers, serially diluted, and plated on antibiotic plates to enumerate bacteria. The percentage of translocated bacteria was calculated as the proportion of the inoculum in the bottom chamber. Three independent experiments, each with ≥ 5 replicates, were conducted.

Strains transformed with either pFUS2 (empty vector) or pFUS2[iucA-D] (plasmid complement) were grown in LB supplemented with 10 µg/mL gentamicin overnight, and subcultured 1:100 into LB supplemented with 200 µM DIP and 5 µg/mL gentamicin, and grown to midlog. Afterwards, the strains OD₆₀₀ were adjusted to 0.5, and resuspended in FBS-free DMEM supplemented with 5 µg/mL gentamicin. Cells were also washed and replaced with media supplemented with 5 µg/mL gentamicin, prior to the start of the assay.

To examine tight junctions (TJs), Transepithelial Electrical Resistance (TEER) measurements were collected using a Millicell ERS-2-Volt-ohmmeter. TEER value >300 Ω/cm2 for the Caco-2 polarized monolayer was indicative of being fully differentiated. Only monolayers with baseline TEER measurements > 800 Ω/cm2 were utilized for these experiments. Assays were conducted as described above, and TEER measurements were conducted prior to infecting (T0) and at various time points post-infection (30 min, 1-, 2-, 4- and 6-hours). Ethylene glycol-bis(β-aminoethyl ether)-N,N,N′,N′-tetraacetic acid (EGTA; 10 mM), known to disrupt TJs [99], and DMEM-only were used as controls. Five independent experiments, each with ≥ 4 replicates, were conducted.

Chrome azurol S (CAS) agar plate assay

The Chrome azurol S (CAS) agar plates were prepared as previously described [100,101] to detect siderophore production. Briefly, 100 mL of CAS dye was prepared by dissolving 10 mM FeCl₃•6H₂O in 10 mL of 10 mM HCl and combined with 50 mL of 2 mM CAS dye in H₂O. The 60 mL mixture was subsequently combined with 40 mL of 5 mM Hexadecyltrimethylammonium bromide (HD^TMA). Separately, 500 mL of King Agar B (Sigma-Aldrich) was prepared. Both reagents were autoclaved for 15 min at 121°C. The cooled CAS dye was slowly added, mixed with the molten King Agar B, and poured into petri dishes.

To qualitatively detect siderophore production, overnight cultures of Kpn strains to be tested were grown at 37°C with agitation under aerobic conditions, diluted 1:10 with 1X PBS and 10 µL spotted on CAS agar plates. Plates were incubated for 16 hours at 37°C. An orange halo surrounding bacteria indicated a positive CAS reaction and the presence of siderophores. Halos were measured using ImageJ version 1.52a software by measuring the area of each halo while excluding the bacterial colony (NIH, USA, https://imagej.net/ij/) [102].

For hvKP94 parental and mutant strain, CAS agar plates were incubated for an additional 72 hours at room temperature to ensure accuracy of siderophore production potential. For assays conducted using the pFUS2 (empty vector) or pFUS2[iucA-D] transformed strains, CAS agar plates were incubated for an additional 24 hours at room temperature.

Histology and imaging

C57BL/6J 5–7 week old mice were orally inoculated with 2% Sucrose PBS (vehicle-only control; n = 2), hvKP1 (WT; n = 2), or hvKP1 iucA- isogenic mutant (n = 3). All mice were euthanized 24 hours PI, and the proximal colons (approximately 4 cm) were collected and immediately flushed with 4% paraformaldehyde (PFA). Fecal pellets were carefully removed, and colons were fixed in 4% PFA for 72 hours. Kpn colonization density in the ileum and cecum at 24 hours was determined by plating tissue homogenates on selective antibiotic plates and enumerating CFU counts.

PFA-fixed proximal colon cross-sections were treated with 70% ethanol and embedded in paraffin and sectioned at 5 µm onto glass slides. Slides were dewaxed and stained with hematoxylin and eosin (H&E), using a routine H&E staining protocol. Briefly, H&E stains were performed using the autostainer XL from Leica Biosystems. The slides were stained with Hematoxylin (Epredia, 7211) for 2 mins and Eosin -Y (Epredia, 7111) for 1 min. Clarifier 2 (7402) and Bluing (7111) solutions from Epredia were used to differentiate the reaction. After staining, slides were then dehydrated in graded ethanol, ending in xylene, and coverslipped with Cytoseal 60 (Epredia, 83104). H&E stained slides were digitally imaged in the Aperio AT2 (Leica) using 20x objective. Blinded histopathological scoring of H&E slides was conducted for each sample.

To detect Occludin antigen, chromogenic Immunohistochemistry (IHC) was performed using the Leica Bond RX Autostainer system on paraffin-embedded mouse colon tissues that were sectioned at 5 µm onto positively charged slides. Slides were blocked and then incubated with Occludin Antibody (ab216327, Abcam) at 1:400 for 1h followed by ready to use Novolink Polymer (RE7260-CE, Leica Biosystems) secondary. Antibody detection with 3,3’-diaminobenzidine (DAB) was performed using the Bond Intense R detection system (DS9263). IHC stained slides were digitally imaged in the Aperio AT2 (Leica Microsystems) using 20x objective.

Occludin image analysis was performed using the HALO AI software (Indica Labs, version 4.0). The mucosa (epithelial cells) was separated from the underlying muscle layers using a MiniNet HALO AI tissue classifier. Next, a customized multiplex IHC algorithm was used to detect and quantify the number of cells positive for Occludin. Nuclei segmentation was performed using the AI default pre-trained algorithm with a hematoxylin weight of 1, occludin nuclear detection weight of 0.04, nuclear contrast threshold of 0.053, nuclear range of 11.3 to 144.81 μm, and nuclear segmentation aggressiveness of 0.55. An AI custom membrane segmentation algorithm detected cellular membranes with a maximum cytoplasm radius of 1.98 μm and membrane segmentation aggressiveness of 0.007. Occludin thresholds were set to 0.15, 0.4268, and 1.1161 for low, medium, and strongly staining cells, respectively. At least 42.2% of the membrane needed to be positive for the cell to be considered positive, since Occludin is mainly located at cell-cell junctions. Both the non-mucosal and mucosal tissue regions were analyzed.

Serum survival/killing assay

To assess complement-mediated killing, Kpn strains grown to mid-log in LB supplemented with 200 µM DIP and adjusted to OD₆₀₀ 0.5, were washed, pelleted and resuspended in PBS to a concentration of 10⁹ CFU/mL. In a 96-well plate, 5 μL of each strain (10⁷ CFU) were combined with 45 μL DMEM, and 5 μL of either rabbit serum (Cedarlane, #CL3111), heat-inactivated (56°C for 30 min) rabbit serum (control), or 1x PBS (control). Following a 1-hour incubation at 37°C, bacterial survival was determined via quantitative enumeration of CFU.

Statistical analysis

All statistical analyses were performed using GraphPad Prism 10.2.3 (GraphPad Software, Inc., San Diego, CA). Comparisons between two groups were analyzed using the Mann-Whitney U test. For comparisons among multiple groups, the Kruskal-Wallis test was applied, followed by post hoc Dunn’s multiple comparisons test or a one-way ANOVA, followed by Tukey’s post-hoc test was performed. Competition experiments were analyzed using the Wilcoxon signed-rank test for matched pairs. Kaplan-Meier curves for survival comparisons were analyzed with the Log-rank (Mantel-Cox) test. The association between bacterial load in the cecal and extra-intestinal samples was assessed using simple linear regressions and a correlation matrix with Pearson correlation coefficients (r) at a 95% confidence interval. Associations between categorical variables in the 3x2 contingency table were assessed using the Fisher’s exact test (Freeman-Halton extension). Data used to generate the figures is presented in S4 Table.