For this study, we used Klebsiella pneumoniae strain #8 (UKP8), a clinical strain isolated from a human patient with urinary tract infection. PCR genotyping using primers described elsewhere (Turton et al., 2008) indicated that the strain UKP8 presents the capsular serotype K2 (data not shown). Bacterial cells were routinely grown in Lysogeny Broth (LB; BD, United States) at 37°C with shaking at 200 rpm, or on LB agar plates in static cultures. Bacterial growth was monitored by measuring the optical density of the cultures at a wavelength of 600 nm (O.D._{600 nm}), using the GeneQuant Spectrophotometer (GE Healthcare). For growth under iron-replete and iron-limiting conditions the LB medium was supplemented with FeSO₄ (Sigma-Aldrich) or the iron chelator Dipyridyl (Sigma-Aldrich), respectively.

For coculture assays, we used the human bladder epithelial cell line T24, purchased from the Bank of Cells of Rio de Janeiro (BCRJ, cell line with code number 0231). T24 cells were cultivated at 37°C in an atmosphere of 95% relative humidity at 5% CO₂, in McCoy’s 5A medium (Thermo Fisher Scientific^®) with 10% fetal bovine serum (FBS) (Thermo Fisher Scientific^®) and subcultivated twice a week at a ratio of 1:8.

Bioinformatics analyses were performed to identify putative regulatory binding sites on kpfR promoter. These analyses were conducted on the genomic sequence of kpfR on Klebsiella pneumoniae strain ATCC 700721/MGH 78578 (GenBank accession number CP000647.1). To identify Fur-binding sites on kpfR gene we employed a theoretical approach (Ferraz et al., 2011) previously adapted to K. pneumoniae (Gomes et al., 2018). To identify the putative transcription start site and possible Sigma factor positions on the promoter region of kpfR gene we used the web-based programs Neural Network Promoter Prediction (available at https://www.fruitfly.org/seq_tools/promoter.html) and BPROM-Prediction of Bacterial Promoters (available at http://www.softberry.com/berry.phtml?topic=bprom&group=programs&subgroup=gfindb; Solovyev and Solovyev, 2011), respectively. These predictive analyses were performed on 100 nucleotides upstream to the initial codon of kpfR gene.

To determine whether kpfR cotranscribe with the genes from kpf cluster, reverse-transcription-PCRs were performed using cDNAs synthesized from total RNA extracted from K. pneumoniae and primer pairs spanning kpfR to kpfA and kpfA to kpfD genes (see Supplementary Table S1). The resulting amplicons were analyzed by agarose gel electrophoresis. Primers were designed using Primer3 version 0.4.0 web-program (available at http://bioinfo.ut.ee/primer3-0.4.0/; Rozen and Skaletsky, 2000).

Complementary oligonucleotides containing the sequences of the putative Fur boxes were annealed to form double-stranded DNA probes. These DNA probes were cloned into high copy number pGEM^®-T Easy vector (Promega) and the recombinant vectors were used in Fur titration assay (FURTA) and DNA Electrophoretic Mobility Shift Assay (EMSA), in order to validate the Fur protein interactions with the putative Fur boxes.

FURTA was performed with Escherichia coli strain H1717, which carries the lacZ reporter gene under the control of the Fur-regulated fhuF gene promoter. When transformed with multicopy plasmids cloned with functional Fur box, H1717 strain will appear red on MacConkey lactose agar plates (Lac⁺ phenotype) because the high number of newly introduced Fur boxes will titrate the Fur repressor from fhuF:lacZ fusion, thus releasing the transcription of lacZ. On the other hand, if H1717 strain is transformed with a vector cloned with a non-functional Fur box, the Fur repressor will remain bound to the fhuF promoter region, and the lacZ reporter gene is not expressed, rendering colorless E. coli H1717 colonies on MacConkey lactose agar plates (Lac^– phenotype). FURTA was conducted as previously described (Johnson and Clegg, 2010). E. coli strain H1717 was transformed with pGEM^®-T Easy vector cloned with the DNA probe containing the sequences of the putative Fur box identified on kpfR gene. The transformants were then plated onto MacConkey lactose agar containing 100 μg/mL ampicillin and 100 mM of iron sulfate (FeSO₄) (Sigma). After 18 h of incubation at 37°C, the functionality of the putative Fur box on kpfR was assessed according to the Lac ± phenotype (i.e., the color of the H1717 colonies). E. coli strain H1717 transformed with circular pGEM^®-T Easy vector alone (i.e., with no insert) was used as negative control, whereas H1717 strain transformed with vector cloned with the previously validated Fur box of the K. pneumoniae entC gene (Carpenter et al., 2009) was used as a positive control.

EMSA was conducted essentially as described by Gomes et al. (2018). DNA probes for EMSA were obtained by PCR amplifying the pGEM^®-T Easy vector cloned with the putative Fur identified on kpfR gene, using the universal M13 primers (Gomes et al., 2018). The resulting PCR product of 285 base pairs was then used as probes on the binding reactions, along with the previously purified His-tagged recombinant Fur protein from K. pneumoniae. The negative control consisted of a 254 base pairs DNA fragment without Fur box sequence, obtained by PCR amplification of a pGEM^®-T Easy vector without insert (i.e., vector not cloned with the putative Fur boxes) using M13 primers. Reactions were performed with 100, 250, and 500 ηM of His-Fur protein, previously equilibrated for 10 min on ice in 10 μL reaction volume containing 1× binding buffer (10 mM Tris, 50 mM KCl, 1 mM DTT, pH 7.5), 0.5 mM MgCl₂, 0.5 mM MnSO₄ and 2.5% (v/v) glycerol. Next, 50 ηg of the DNA probes was added and the mixture was incubated for 20 min on ice. EMSA was also performed under divalent cation-free conditions by adding EDTA to a final concentration of 2 mM in the above reaction mixture. Samples were loaded onto a 2% (w/v) agarose gel prepared with 1× Bis-Tris borate buffer containing 0.1 mM MnSO₄. After 30 min of electrophoresis, the gels were stained with ethidium bromide solution (0.5 μg/mL) for 15 min, and the DNA bands were visualized and recorded under a digital photodocumentation system (Biorad).

Knockout of the kpfR gene from K. pneumoniae UKP8 was constructed using the TargeTron Gene Knockout System (Sigma-Aldrich), according to the manufacturer’s instructions. This system is based on site-specific and not random disruption of the gene of interest by insertion of group II intron harboring a kanamycin resistance gene (kan^R).

Firstly, a computer algorithm at TargeTron Design website (Sigma-Aldrich) was used to identify potential target sites for intron insertion on the kpfR coding region. The most efficient target site was selected, based on the lowest E-value predicted by TargeTron Design algorithm (see Supplementary Table S2). The TargeTron Design algorithm also provided the primers used in PCR reactions to mutate (re-target) the RNA segment of the intron (see Supplementary Table S3). The 350 pb PCR product was digested and ligated into chloramphenicol-resistant pACD4K-C vector provided by the manufacturer (Sigma-Aldrich). The resulting recombinant pACD4K-C vector was transformed into Escherichia coli DH5a by heat shock, and clones of the recombinant vector were obtained using Wizard Plus SV Minipreps DNA Purification kit (Promega). The pACD4K-C vector contains a T7 promoter and, therefore, requires T7 RNA Polymerase to express the retargeted intron that will disrupt the target gene. Since Klebsiella pneumonia does not express T7 RNA Polymerase, it was necessary to use the ampicillin-resistant pAR1219 vector (Sigma-Aldrich) in co-transformations because this vector contains the T7 RNA Polymerase gene under the control of the IPTG-inducible lac UV5 promoter. Chemically competent K. pneumoniae UKP8 cells were co-transformed with pAR1219 and the pACD4K-C recombinant vectors and cultured overnight in LB containing 100 μg/mL of ampicillin and 25 μg/mL of chloramphenicol. On the next day, the culture was diluted 1:50 in fresh LB containing ampicillin and chloramphenicol, and cell growth was monitored until they reached O.D._{600 nm} of 0.2. At this point, to induce intron expression and insertion, 1 mM (final concentration) of IPTG was added to the medium and the culture was incubated at 30°C for 30 min. After induction, cells were harvested by centrifugation at 14,490 g for 1 min, resuspended in antibiotic-free LB broth, and incubated again for 1 h at 30°C. The cells were then plated on LB agar supplemented with 25 μg/mL kanamycin and grown at 30°C for 2–3 days. Knockout bacteria were selected from kanamycin-resistant colonies and the insertion of the intron RNA harboring kan^R gene was confirmed by PCR using primers flanking the target insertion site in the coding region of the kpfR gene (Supplementary Table S1). Since the knockout is based on the insertion of kan^R gene on kpfR, the K. pneumoniae UKP8 mutant strain was renamed kpfR::kan^R.

A complemented strain, named kpfR::kancompR, was obtained by introducing the kpfR gene back into the mutant strain. For this, a DNA fragment comprising the entire coding region of kpfR plus approximately 550 bp of 3’ and 5’ flanking regions (Supplementary Table S4) was inserted on pCR2.1-TOPO vector (Invitrogen) previously cloned with erythromycin-resistance gene. Chemically competent kpfR::kan^R was transformed with the recombinant vector and plated on LB agar supplemented with 50 μg/mL erythromycin. Complemented strains were recovery by screening erythromycin-resistant colonies.

Initially, the growth of wild-type UKP8 and mutant kpfR::kan^R strains was assessed to investigate whether the lack of KpfR compromises bacterial growth. Strains were separately inoculated into LB medium and grown until saturation (overnight) at 37°C while shaking. The next day, the culture was diluted 1:200 in fresh LB and the bacterial growth was monitored every 15 min by measuring the O.D._{600 nm}. Growth curves were constructed by plotting the O.D._{600 nm} values against time.

Ten microliters of saturated cultures of UKP8, kpfR::kan^R, and kpfR::kancompR were plated on blood agar (Trypticase Soy Agar plates supplemented with 5% sheep blood) for 18 h at 37°C to check for macroscopic differences in each strain’s morphology.

Motility assays were carried out in triplicate for each K. pneumoniae strains, according to protocol described elsewhere with minor modifications (Losensky et al., 2015). In brief, 10 μL of bacterial cultures standardized at O.D._{600 nm} of 0.4 was inoculated in the center of soft LB agar plates (prepared with 0.8% bacteriological agar) and incubated at 37°C for 120 h. After this, the extension of the colonies diameter was measured in cm.

Transmission Electron Microscopy (TEM) was carried out for visualization of fimbriae production by the K. pneumoniae strains. Bacterial cells were grown statically on LB agar plates at 37°C for 18 h and resuspended in 500 μL of PBS buffer. A droplet of the cell suspension was spotted on Formvar coated carbon-reinforced copper grid (400 mesh) and incubated at room temperature for 2 min. Grids were negatively stained using a drop of 1.25% phosphotungstic acid pH 6.5 (Sigma-Aldrich) filter sterilized through a 0.22 μm pore-size filter membranes (Millipore). The staining lasted 15 s and then the grids were air-dried on a piece of filter paper. Samples were examined and electron micrographs were obtained with a Zeiss LEO 906 Transmission Electron Microscope operated at 60 kV.

Biofilm formation assays were conducted in polyvinylchloride (PVC) microplates, following a protocol described elsewhere with minor modifications (O’Toole and Kolter, 1998). This protocol is based on the ability of bacteria to form biofilm on PVC microplates and on the ability of crystal violet to stain bacterial cells but not PVC. Strains were grown until saturation at 37°C under shaking. Bacterial cells from saturated cultures were harvested by centrifugation and resuspended in LB broth to a final concentration of 10⁶ cells/mL. 5 μL of this suspension were added on flat-bottom 96-well PVC microplates containing 150 μL of LB and incubated at 37°C for 1, 5, and 10 h under static conditions. After incubation, the culture medium was removed and the wells were gently washed twice with sterile water to remove unattached cells. Next, 25 μL of 1% crystal violet was applied in the wells and the plates were incubated for 15 min at room temperature. After incubation, the excess of dye was removed and the wells were washed with deionized water. Then, the remaining crystal violet staining of the adherent cells was solubilized with absolute ethanol and the absorbance of the ethanol containing the eluted dye was measured at O.D._{600 nm} in a spectrophotometer. The biofilm formation assays were repeated at least three times.

Yeast agglutination assays were performed to investigate the expression of fimbriae by the K. pneumoniae strains. Assays were carried out in triplicate for each K. pneumoniae strain and were conducted on Kline concavity slides, according to protocol previously described (Roe et al., 2001; Schembri et al., 2004). Bacterial strains grown on LB agar plates up to 72 h were resuspended in PBS buffer and the concentration was standardized at O.D._{600 nm} of 0.6. Bacteria were then mixed with an equal volume of 5% (wt/vol) suspension of Saccharomyces cerevisiae cells (Sigma-Aldrich) prepared in PBS buffer. The time required for agglutination and its intensity were documented. The agglutination of the yeast cells is specifically mediated by type 1 fimbriae since these fimbriae have great affinity for mannose, a highly abundant residues on yeast cell-surface. Therefore, the assays were also performed in the presence of 5% D-(+)-Mannose (Sigma-Aldrich) to confirm if the agglutination was indeed mediated by the type 1 fimbriae.

To visualize capsular polysaccharide (CPS) produced by UKP8, kpfR::kan^R, and kpfR::kancompR, strains were cultured under the same growth conditions, stained with India ink, and analyzed under optical microscopy. The presence of capsule is indicated by a negative staining area around the bacteria. CPS production was quantified by extracting and measuring glucuronic acid, a significant component of the K. pneumoniae capsule, using a colorimetric assay previously described (Lin et al., 2012; Ares et al., 2016). Briefly, 500 μL of K. pneumoniae cultures were mixed with 100 μL of 1% Zwittergent 3–14 detergent (Sigma-Aldrich) in 100 mM citric acid (pH 2.0) and then incubated for 20 min at 50°C. After centrifugation for 2 min at 14,490 g, 250 μL of the supernatants were transferred into new tubes, 1 mL of cold absolute ethanol was added, and the mixtures were incubated for 20 min at 4°C for CPS precipitation. After centrifugation at 14,490 g for 2 min, the supernatants were discarded and the pellets were allowed to dry. The dried pellets were dissolved in 200 μL of distilled water and then 1,200 mL of 12.5 mM tetraborate in concentrated H₂SO₄ were added. Next, the mixtures were vigorously vortexed, heated in a boiling-water bath for 5 min, and cooled down, before the addition of 20 μL of 0.15% 3-hydroxydiphenol (Sigma-Aldrich) prepared in 0.5% NaOH. Then, the absorbance was measured in a spectrophotometer at a wavelength of 520 nm. The glucuronic acid concentration in each sample was determined from a standard curve of glucuronic acid (Sigma-Aldrich) and expressed in micrograms/10⁹ CFU. Quantification of CPS production was performed from five independent cultures of the strains.

Coculture assays were performed as previously described (Elsinghorst, 1994; Oelschlaeger and Tall, 1997), with some modifications. Saturated overnight cultures of UKP8 and kpfR::kan^R were inoculated on fresh LB medium and incubated at 37°C with shaking until they reached the mid-logarithmic growth phase (O.D._{600 nm} of 0.4). Bacteria were added to an 80% confluent T24 cell monolayer seeded into 12-well plates at a multiplicity of infection (MOI) of 200 bacteria per host cell, and the infected monolayers were incubated at 37°C for three different periods, specific for each assay. For adhesion assay, infected T24 cells were incubated for 30 min, washed three times with phosphate-buffered saline (PBS buffer) to remove non-adhered bacteria, and lysed by adding 0.1% Triton X-100 diluted in PBS. Serial dilutions of the lysate were plated on blood agar and incubated overnight at 37°C for counting bacterial colony-forming units (CFUs). For invasion assay, infected T24 cells were initially incubated for 4 h to allow invasion of the bacterial strains into T24 cells. Plates were washed once with PBS buffer and incubated again for an additional 1 h in fresh culture medium containing the antibiotic gentamicin (25 μg/mL) to eliminate extracellular bacteria (both planktonic and adhered). The plasma membrane of host cells is impermeable to gentamicin. Therefore, after treatment with this antibiotic, only extracellular bacteria will be eliminated, while those present in the intracellular environment will be preserved. After treatment with gentamicin, the T24 cells were washed, lysed with Triton X-100 and serial dilutions of the lysate were plated on blood agar for counting of CFUs. For intracellular replication assay, infected T24 cells were initially incubated for 4 h to allow invasion of the bacterial strains. Plates were washed once with PBS buffer and incubated again for 24 h in fresh culture medium containing gentamicin at 10 μg/mL. Then, the T24 cells were washed, lysed with Triton X-100 and serial dilutions of the lysate were plated on blood agar for CFUs counting. The integrity of the T24 cells during the coculture assays was assessed by the standard exclusion protocol of trypan blue 0.4% (Thermo Fisher Scientific^®).

For mouse urinary tract infection analyzes were used 6–8 weeks old female BALB/c mice (CEMIB, UNICAMP, Brazil). Infections were carried out by transurethral inoculation as previously described (Thai et al., 2010; Reis et al., 2011). UKP8 and kpfR::kan^R strains were grown on LB agar plates at 37°C for 18 h and resuspended at a concentration of ∼5 × 10⁸ CFU/mL. Fifty microliters of the bacterial suspension were inoculated by transurethral catheterization in mice anesthetized intraperitoneally with ketamine and xylazine. Mice were euthanized at the indicated times, and urine and bladders were aseptically collected and processed for histology and CFU titration. Serial dilutions of urine and homogenized bladders were plated on blood agar and incubated overnight at 37°C for counting of CFUs. Whole bladders were fixed in 10% neutral buffered formalin, embedded in paraffin, and 5 μm-thick sections were stained with hematoxylin and eosin (H&E).

Mice experimental procedures were conducted in accordance with guidelines of the Brazilian College for Animal Experimentation (COBEA) and received prior approval by the Ethics Committee on the Use of Animals in Research of Universidade São Francisco (CIAEP/CONCEA No. 01.0226.2014, protocol number 004.04.2015).

Gene expression analysis was performed using Reverse Transcription Quantitative real-time Polymerase Chain Reaction (RT-qPCR). The expression of kpfR under iron-replete and iron-limiting conditions was investigated as previously described (Gomes et al., 2018). Briefly, K. pneumoniae strains were grown in LB broth at 37°C with shaking at 200 rpm until they reached the mid-logarithmic growth phase (O.D._{600 nm} of 0.4). At this point, ferrous iron (100 μM of FeSO₄, final concentration) or an iron chelator (100 μM of 2,2’-Dipyridyl, final concentration) were added and the cells were incubated for 1 h. After 1 h incubation, bacterial cells were harvested by centrifugation and the pellets were resuspended in RNAprotect^® Bacteria Reagent for RNA stabilization. All culture conditions were performed at least twice. Control condition consisted of K. pneumoniae cells grown in LB medium without supplements.

The expression of fimbrial genes and galF, the first gene of the capsular polysaccharide synthesis (cps) gene cluster, were also assessed on UKP8 and kpfR::kan^R strains grown on blood agar plate. For this, bacterial colonies from each strain were individually collected from the blood agar plates, and the cell pellets were resuspended in RNAprotect^® Bacteria Reagent until the moment of the RNA extraction. Total RNA from the cell pellets was extracted using RNeasy Protect Bacteria Mini Kit (Qiagen), following the manufacturer’s protocol. An on-column DNase digestion with the RNase-free DNase Set (Qiagen) was performed to remove genomic DNA contamination in RNA samples.

To assess gene expression of the bacterial strains during coculture with T24 bladder epithelial cells, the adhesion and invasion assays were conducted as previously described in section “Materials and Methods,” the only exception that the T24 cells were seeded in 6-well plates. After the incubation periods, extracellular bacteria were removed by washing with PBS buffer and T24 cells were collected for the extraction of bacterial total RNA. RNA was extracted using Max^TM Bacterial RNA Isolation kit (Ambion), following the manufacturer’s instructions. MICROBEnrich^TM kit (Ambion) was applied, in order to promote the selective removal of total RNA from T24 cells, preserving bacterial RNA. Total RNA extracted was further treated with DNAse.

After treatment with DNAse, 0.5–1 μg of total bacterial RNA was used for cDNA synthesis by using the ThermoScript^TM RT-PCR System for First-Strand cDNA Synthesis Kit (Invitrogen) according to the manufacturer’s instructions. The synthesized cDNAs were used in RT-qPCR reactions done in triplicates using the Platinum^® SYBR^® Green qPCR SuperMix-UDG Kit (Invitrogen) on 7300 Real-Time PCR System equipment (Applied Biosystems). Data were normalized using rho and recA as endogenous genes, which encode transcription termination factor and recombinase A, respectively (Song and Abraham, 2008). The relative expression levels of the selected genes were calculated by the comparative critical threshold (2^–ΔΔCT) method (Livak and Schmittgen, 2001). GraphPad Prims 7.00 (GraphPad Software, Inc.) was used for the statistical analyses. Differences on the expression levels were evaluated by Student’s t-test, and differences with p ≤ 0.05 were considered statistically significant. Primers used on RT-qPCR reactions are listed in Supplementary Table S5 and were designed using Primer3 version 4.1.0 web-program.

The kpf gene cluster comprises 4 genes designated as follow: kpfA, encoding the major pilin subunit, kpfB, encoding a chaperone, kpfC, encoding an usher, and kpfD, encoding an adhesin. The first gene adjacent to kpf cluster at 5’ extremity encodes a putative helix-turn-helix transcriptional regulator. To determine whether this adjacent gene belongs to kpf cluster, reverse-transcription-PCRs were performed using cDNA and primer pairs spanning the entire cluster and confirmed that kpf cluster and the adjacent gene are co-transcribed as a single polycistronic transcript (Figure 1A). Since the adjacent gene encodes a putative helix-turn-helix transcriptional regulator, and following Wu et al. (2010) nomenclature, we designated it kpfR gene.

To better understand the expression regulation of kpf, bioinformatic analyses were employed to identify regulatory domains in the promoter region of the cluster. Thus, a putative Fur box sequence was identified 100 nucleotides downstream to start codon of the kpfR gene (Figure 1B). The long-distance between the Fur box and the promoter region of the kpf cluster led us to investigate additional regulatory elements on the adjacent sequences. As indicated in Figure 1B, a putative binding site of the DNA bending protein IHF (integration host factor) was identified at the intervening region between the promoter and the Fur box. This putative site closely resembles to the consensus recognition sequence of IHF, but its functionality was not validated in the present study.

The putative Fur box identified inside the coding region of kpfR was validated by Fur Titration Assay (FURTA) and DNA Electrophoretic Mobility Shift Assay (EMSA). Figure 1C shows the FURTA-positive control (Lac⁺, red colonies), the FURTA-negative control (Lac^–, colorless colonies) and the validation of the Fur box on kpfR, indicated by the red colonies. This result confirms that Fur regulator from E. coli H1717 was able to bind in vivo to the cloned putative Fur box from kpfR of K. pneumoniae, rendering the colonies with red color. Next, EMSA was performed to verify the direct interaction of K. pneumoniae Fur protein on the putative Fur box found on kpfR. As shown on Figure 1D, a mobility shift of the DNA probes containing the putative Fur box was observed when increasing concentrations of K. pneumoniae purified His-Fur protein was incubated with these probes. Besides, the addition of the divalent cations chelator EDTA abolished the mobility shift of the probe, indicating that divalent cations are required for Fur interaction. No mobility shift is observed when K. pneumoniae Fur protein is incubated with a DNA probe without Fur box sequence.

Taken together, our results confirm that Fur protein recognizes and binds to the Fur box on kpfR of K. pneumoniae, indicating that the expression of kpfR seems to be modulated by Fur transcriptional regulator in an iron-dependent manner. To further investigate how Fur modulates the expression of kpfR, RT-qPCR analyses were performed on K. pneumoniae cells kept under iron-replete and iron-limited conditions. As displayed on Figure 1E, the iron-replete condition presents no changes in the expression pattern of kpfR gene when compared to the control condition. On the other hand, kpfR is repressed under condition of iron scarcity. These results show that Fur modulates the expression of kpfR gene according to the availability of iron in the culture medium. Thus, kpfR belongs to the Fur regulon of K. pneumoniae.

To evaluate the role of KpfR as regulator of morphology and virulence of Klebsiella pneumoniae, a mutant (kpfR::kan^R) was generated (as described in section “Materials and Methods”). The mutant strain exhibits prolonged growth at the lag phase, leading to delayed entry into the logarithmic (log) phase of growth (see Supplementary Figure S1). Nonetheless, once reaching the log phase, the mutant strain displayed a growth rate similar to that observed in the wild-type strain. Slight differences in growth patterns were observed among the UKP8 wild-type strain and the complemented kpfR::kancompR strain; i.e., the mutant strain transformed with a kpfR-carrying vector.

Macroscopic morphologies of wild-type UKP8 and mutant kpfR::kan^R strains were investigated on blood agar plates. As shown on Figure 2A, the wild-type strain exhibited circular shape, high elevation, entire margin, and smooth bright surface. On the other hand, the mutant kpfR::kan^R strain presented irregular shape, flat elevation, lobed margin, and rough matte surface. In addition, kpfR::kan^R colonies exhibited a slight dispersion on the plates. The morphological features observed on the mutant strain can be attributed to the lack of KpfR, because the pattern observed on the wild-type strain was fully reestablished on the complemented mutant strain (kpfR::kancompR).

The spreading behavior of kpfR::kan^R colonies on blood agar plates prompted us to perform motility assays on soft agar plates. As displayed on Figure 2B, the wild-type UKP8 strains presented a slight and linear radial migration on soft agar plates. On the other hand, kpfR::kan^R cells exhibited a more prominent and asymmetric dispersion on the surface of the plates. The complemented mutant strain was partially restored since it had a reduced migration extension when compared to both wild-type and mutant strains. The extent of the area in which the wild-type, mutant, and complemented strains migrated on the plates was measured, and confirmed that the mutant strain had a greater dispersion when compared to UKP8 and kpfR::kancompR strains (Figure 2C). These results indicate that kpfR disruption provided a sort of rudimentary motility to the mutant bacteria.

Since loss of kpfR rendered K. pneumoniae a prominent dispersion on agar plates, we evaluated the cell surface morphology of the wild-type, mutant, and complemented strains by Transmission Electron Microscopy (TEM). While UKP8 strain appears with no surface appendages, the cell surface of the kpfR::kan^R cells was covered by numerous appendages similar to fimbriae or pili that are long and flexible with a varied size larger than 1 μm (Figure 3A). These numerous appendages were no longer seen on the surface of the complemented mutant strain (kpfR::kancompR).

The presence of numerous appendages on kpfR::kan^R cells led us to investigate the expression of fimbrial genes on UKP8 and kpfR::kan^R cells grown on blood agar plate. Disruption of kpfR resulted in up-regulation of fimA, from type 1 fimbrial gene cluster fim, and kpfR and kpfA genes, from kpf gene cluster (Figure 3B). kpfR and kpfA presented a similar expression pattern, confirming the polycistronic transcription of these genes, as previously shown. Moreover, the up-regulation of kpfR in the cells with disrupted kpfR suggests that KpfR autoregulates its own expression. On the other hand, lack of KpfR regulator resulted in down-regulation of ecpA gene, from the ecp cluster, whereas mrkA gene, from type 3 fimbrial gene cluster mrk, presented unchanged expression pattern between wild-type and mutant strains.

To investigate the phenotypic effects of increased type 1-like fimbriae production due to KpfR absence, biofilm formation, and yeast agglutination assays were performed with the UKP8, kpfR::kan^R, and kpfR::kancompR strains. Biofilm formation by kpfR::kan^R cells was significantly superior than UKP8 after 3, 5, and 10 h of incubation (Figure 3C). The complemented kpfR::kancompR strain failed to restore the biofilm formation presented by the wild-type UKP8, since exhibited a phenotype similar to the mutant kpfR::kan^R strain. Yeast agglutination assays allow specific detection of type 1 fimbriae because type 1 fimbrial adhesins have a great affinity for mannose-containing receptors on the yeast cell surface. UKP8 showed no yeast agglutination, whereas the kpfR::kan^R promptly agglutinated the yeast cells (Figure 3D). The lack of agglutination observed on UKP8 was fully restored on the complemented kpfR::kancompR strain. To determine if the agglutination pattern exhibited by the mutant strain was due to type 1 fimbriae expression, the assay was also performed in the presence of mannose prior to the addition of the yeast cells. In the presence of mannose, the mutant strain loses the ability to agglutinate yeast cells. These results reveal that the observed agglutination of yeast cells was indeed mediated by type 1 fimbriae.

As the lack of KpfR regulator resulted in increased expression of type 1-like fimbriae, we decided to investigate the production of capsular polysaccharide, another important virulence factor of the bacterial cell surface. Negative staining with India ink revealed no visible capsule in the kpfR:::kan^R mutant strain as compared to the UKP8 wild-type strain (Figure 4A), while this phenotype was partially restored on the complemented strain. In addition, quantification of glucuronic acid revealed that the mutant strain presents 10-fold decrease in capsule production compared to that in the wild-type strain (Figure 4B). Capsule production by the complemented strain was partially restored to the levels observed in the wild-type strain. We next investigated the expression of the cps gene cluster on the wild-type and the mutant strains by determining the expression levels of galF using RT-qPCR analysis. galF is the first gene of the cps cluster and encodes an enzyme dedicated to the synthesis of sugar nucleotides, which are the precursors that will compose the repeating-units of the capsular polysaccharide. Consistent with the less capsule production by the kpfR::kan^R mutant strain, RT-qPCR analysis revealed down-regulation of the capsule-producing gene galF on kpfR::kan^R in comparison with the expression on UKP8 wild-type strain (Figure 4C).

Considering fimbriae important mediators of bacterial adhesion and internalization into host cells, we investigated the role of kpfR gene in host-pathogen interactions by assessing the adhesion, invasion, and intracellular replication of UKP8 and kpfR::kan^R strains on human bladder epithelial cell line T24. The data presented on Figure 5A indicate that the kpfR::kan^R strain showed greater adhesion to T24 bladder cells than the wild-type UKP8 strain. Moreover, only the mutant strain was able to invade and replicate within T24 cells (Figure 5A). No significant changes in T24 cells viability were observed at a multiplicity of infection (MOI) of 200 bacteria:1 T24 cell and incubation of coculture for 24 h (data not shown). According to the results, the lack of KpfR repressor improved the ability of the mutant strain to adhere, invade and even replicate intracellularly in T24 bladder cells. We next evaluated the expression of kpfR and kpfA (cluster kpf) and fimA (cluster fim) in UKP8 and kpfR::kan^R strains during adhesion to the T24 cells. As shown on Figure 5B, all three genes had the expression significantly induced in kpfR::kan^R compared to UKP8.

To further evaluate the in vivo role of KpfR repressor in Klebsiella pneumoniae virulence, a mouse urinary tract infection model was used to compare the ability of UKP8 and kpfR::kan^R strains to colonize the bladder of the animals (Figure 5C). At 6 hpi, there were no differences in UKP8 and kpfR::kan^R load in mice urine and bladder. At 24 hpi, UKP8 was recovered from bladder tissue and urine, while kpfR::kan^R was not present on these samples. At 48 hpi, UKP8 and kpfR::kan^R were no longer found in urine, whereas only UKP8 was still present in mice bladder tissue.

Aliquots of urine were analyzed under an optical microscope (data not shown) and revealed red blood cells, leukocytes and scaly epithelium in urine of the animals infected with both UKP8 and kpfR::kan^R strains at 6 hpi. At 24 hpi, a larger population of leukocytes and bacteria were observed only in the urine samples of mice infected with the wild-type strain. Urine of mice infected with kpfR::kan^R during 24 h presented neither erythrocytes nor leukocytes.

The colonization experiments suggested that the lack of KpfR repressor confers a disadvantage on kpfR::kan^R strain. To characterize these effects further, histological analyses were performed to obtain images of mouse bladders infected with the wild-type and kpfR::kan^R mutant strains. As shown on Figure 5D, an inflammatory infiltrate consisting mainly of neutrophils and hyperemia of the bladder blood vessels was observed 6 hpi with both strains. At 24 hpi, the hyperemia was more prominent in the bladder tissue infected with UKP8 than kpfR::kan^R, and an abundant migration of inflammatory cells was noticed only with UKP8 at 48 hpi (Figure 5D). In addition, only wild-type strain was able to form intracellular bacterial communities (IBCs) in the urothelium at 24 and 48 hpi (Figure 5E). Taken together, the data show successful bladder colonization and formation of IBC in urothelial cells by the wild-type UKP8 strain, while the kpfR::kan^R mutant strain loses resistance and ability to survive in the host.