Background

Extraintestinal pathogenic Escherichia coli (ExPEC) is one of the most common bacterial pathogens causing bacteremia, neonatal meningitis, and urinary tract (UT) infections (UTIs) [1]. The diseases caused by ExPEC have resulted in substantial morbidity, mortality, and healthcare costs [1, 2]. Antibiotic therapy is the traditional way to treat E. coli infections. However, the rapid emergence of antibiotic-resistant strains has become a serious problem in managing bacterial infections because of the shortage of novel and effective antibiotics [3]. Accordingly, new antimicrobial strategies against E. coli-associated infections are urgently needed. As bacterial factors required for maintaining the virulence of ExPEC are potential antimicrobial targets, identifying such factors and understanding how they contribute to infections would facilitate the development of novel treatment strategies.

The E. coli periplasmic protease Prc is required for ExPEC to cause a high level of bacteremia [4] since ExPEC lacking Prc displays enhanced sensitive to complement-mediated serum killing and thus is defective in survival in the host bloodstream [4]. In addition to that in ExPEC, Prc homologs in other pathogenic bacteria have also been shown to contribute to bacterial pathogenesis. For example, a prc mutant of Salmonella typhimurium exhibits a diminished ability to survive in murine macrophages and attenuated virulence in mice [5]. Disruption of the Prc-homologous protein CtpA in the animal pathogens Brucella suis and Burkholderia mallei decreases the abilities of these bacteria to survive in murine macrophages [6, 7]. Mutation of prc in the plant pathogen genus Xanthomonas results in decreased virulence, biofilm production, and resistance to environmental stresses [8, 9]. In this study, we further demonstrate the novel pathogenic role of Prc in ExPEC UTIs.

The mechanism of how the Prc protease and its homologs contribute to bacterial virulence remains to be elucidated. Our previous study has shown that deletion of prc in the ExPEC strain RS218, which is associated with neonatal meningitis, significantly changed the protein profiles in the outer membrane (OM) fraction [4]. The altered protein expression in the OM fraction may contribute to the defective ability to cause infections since OM-associated proteins (OMPs) are the major factors involved in bacterium-host interactions and play key roles in maintaining the integrity of the OM, which is the main bacterial structure for sensing and coping with the harsh host environment during infections [10].

Flagella are the protein structures associated with the bacterial OM that mediate bacterial motility [11]. Flagella of ExPEC have been shown to contribute to the pathogenesis of UTIs because these structures enable the bacteria to disseminate, and they facilitate colonization and ascension of the UT [12,13,14,15]. The process of flagellum biogenesis is regulated by the flagellar regulon organized in a three-tier hierarchy [16]. Three flagellar genes, flhD, flhC, and fliA, are central for the hierarchical expression of this regulon. At the top of this hierarchy (class 1) are the master operon genes, flhDC. Their gene products, FlhD and FlhC, assemble into a heterohexamer (FlhD4C2) [17] that acts as an essential transcription activator of the class 2 genes. Class 2 genes encode the flagellum-specific sigma factor σ28 (FliA), the flagellar basal body and hook proteins, etc. Class 3 genes encode the subunit of the flagellar filament (FliC), stator components of the flagellar motor, as well as the chemotaxis pathway.

The flagellar regulon is highly regulated by environmental cues, such as osmolality, nutrients, cell density, and temperature [18]. E. coli cells perceive and respond to such external environmental stimuli through extracytoplasmic stress signaling systems (ESSSs) whose activation triggers transcriptional reprogramming, allowing the bacteria to cope with the corresponding external conditions. Two-component signal transduction systems (2CSTSs) are among the members of ESSSs. The activation of some 2CSTSs, including RcsCDB, CpxA-CpxR, EnvZ-OmpR, and QseB-QseC, has been shown to suppress the expression of the flagellar regulon, [19,20,21,22,23,24]. In E. coli, the prototypical 2CSTSs consist of an inner membrane-bound sensor kinase and a DNA-binging cytoplasmic response regulator. In response to specific stimuli, the sensor kinase is autophosphorylated at a conserved histidine residue. Then, the phosphoryl group is transferred to a conserved aspartate in the cognate response regulator. Finally, the phosphorylated response regulator up- or downregulates the transcription of target genes to induce cellular responses to external signals. In addition to 2CSTSs, the alternative sigma factor σE governs an extracytoplasmic signaling pathway that responds to heat-shock stress [25, 26]. Although no study has demonstrated whether the activation of this heat-shock response system suppresses flagellum expression, it is known that bacterial motility is suppressed under high environmental temperatures, suggesting that the σE system may contribute to the regulation of the flagellar regulon [27]. In unstressed bacteria, σE is sequestered in the cytoplasmic side of the inner membrane by the antisigma factor RseA, which is an inner membrane-spanning protein with a periplasmic-exposed C-terminus and a cytoplasmic-exposed N-terminus [28,29,30]. Under stress, RseA is proteolytically degraded by the sequential action of the periplasmic and cytoplasmic proteases DegS and RseP, resulting in release of σE into the cytoplasm, in which this sigma factor can associate with the core enzyme of RNA polymerase to allow σE-regulated gene transcription [31,32,33,34,35,36]. The E. coli envelope, which is composed of the OM, inner membrane (IM), periplasm, and peptidoglycan mesh [37] is the frontline of bacterial interaction with the external environment. Alteration of the envelope components has been shown to be able to activate ESSSs, similar to environmental stresses. It remains unclear whether the altered protein profile of the OM fraction caused by Prc deficiency could activate the signaling systems.

As a periplasmic protease, Prc exerts its biological function through proteolytic regulation of its substrates, which are supposed to be located in or partially exposed to the periplasmic space. The attenuated virulence resulting from Prc deficiency in bacterial pathogens may be a consequence of the substrate dysregulation caused by the loss of proteolytic control. In E. coli, the periplasm-exposed OM protein Spr has been shown to be a substrate of the Prc protease. Spr is a peptidoglycan hydrolase. Singh et al. have shown that deletion of prc causes Spr accumulation in E. coli cells [38]. The protein accumulation contributes to the mutant’s growth defect under low osmolarity at 42 °C [38].

In this study, we found that the Prc protease of ExPEC is necessary for maintaining intact bacterial motility that is important for UT colonization. The underlying mechanism was shown to be involved in the activation of ESSSs and the intracellular level of the Prc substrate Spr.

Methods

Bacterial strains, plasmids and growth conditions

The bacterial strains and plasmids used in this study are shown in Table 1. Bacteria were grown in Luria Bertani (LB) broth containing 1% tryptone, 0.5% NaCl and 1% yeast extract at 37 °C overnight (approximately 16 h) and were stored in LB with a final concentration of 15% glycerol at − 80 °C.

Table 1 E. coli strains and plasmids used in this study
Full size table

Construction of mutants and plasmids

The ExPEC mutants were constructed using polymerase chain reaction (PCR) product-based λ Red recombination, as described previously [42]. The plasmids pCA3 × Flag [43] and pKD3 [42] served as templates for synthesizing the 3 × Flag- and chloramphenicol resistance cassette-encoding sequences by PCR, respectively. The primers used for mutant construction are shown in Additional file 1: Table S1.

The plasmids producing the Prc-S430A or Prc-K455A mutation proteins were generated by site-directed mutagenesis using the corresponding primer (Additional file 1: Table S1). The complementary and overexpression plasmids newly constructed in the study were created by cloning the indicated PCR-amplified gene fragments into the corresponding plasmid vectors. The primers for amplification of the gene fragments are shown at Additional file 1: Table S1.

Mouse model of urinary tract infection (UTI)

The animal UTI studies were performed as described previously [12], with some modification. For each experiment, 2 ExPEC strains were mixed at a ratio of 1:1. Eight-week-old female C3H/HeN mice were anesthetized and transurethrally inoculated with a 50-μl bacterial suspension (1 × 108 colony-forming unit, CFU) per mouse using a sterile polyethylene catheter connected to an infusion pump (Harvard Apparatus, Holliston, MA, USA) with a flow rate of 100 μl/min. Subsequently, 48 h post-infection, mice were sacrificed, and their bladders and kidneys were collected, weighed, and homogenized in sterile culture tubes containing 3 ml of normal saline. Bacterial counts were differentiated and determined by plating the homogenates onto LB agar plates containing IPTG and X-gal. The strains with and without lacZ showed blue and white colonies on the plates, respectively.

Liquid chromatography-tandem mass spectrometry (LC/MS/MS) analysis of bacterial proteins

To prepare bacterial proteins for LC/MS/MS analysis, 16-h cultures of WT-RS218 and Δprc-RS218 (three independent cultures for each strain) were harvested and subjected to French press at 8000 lb./in2. The resulting bacterial lysates were subjected to 12.5% SDS-PAGE to separate the proteins in the samples. The gel lane of each sample was cut into 5 slices. The gel slices were subjected to in-gel digestion with trypsin, followed by protein identification with the Thermo LTQ-Orbitrap Velos system. The MS/MS spectra were searched against Escherichia coli SwissProt 2014_08 (546,238 sequences; 194,363,168 residues) using Sequest (version 27, rev 12), which is part of the BioWorks 3.3 data analysis package (Thermo Fisher, San Jose, CA, USA). Subsequently, protein identifications with 2 peptides in at least one of the samples were retained. The proteins that are defined to be located on or associated with the outer membrane (OM) based on the EcoCyc database (http://ecocyc.org/), showed at least a 2-fold change with statistical significance between WT-RS218 and Δprc-RS218 were identified (Table 2).

Table 2 Identification of altered OMPs by liquid chromatography-tandem mass spectrometry
Full size table

Western blot analysis

The protein levels in the OM fractions or bacterial lysates were determined by western blot analyses. Preparation of the protein samples was performed as described previously [4]. The primary antibodies utilized to detect FliC, Prc, and OmpA were rabbit antisera against FliC (anti-H7, Becton Dickinson, Sparks, MD, USA) and Prc, and a mouse anti-OmpA antiserum. The recombinant proteins fused with HA and Flag tags were detected with a mouse anti-HA antibody and rabbit anti-Flag antibody, respectively (Sigma-Aldrich, St. Louis, MO, USA).

Motility assay

Bacterial strains were stab inoculated onto 0.3% agar plates and incubated at 37 °C for 10 h [44]. The diameter of motility was measured and is shown in the quantified figure.

RNA isolation

Total RNA was extracted from 16-h cultures of bacteria using the RNeasy Mini kit (QIAGEN, Hilden, Germany) according to the manufacturer’s instructions. To remove contaminating DNA, the resulting RNA was incubated with DNase I (Roche Applied Science, Mannheim, Germany) at 37 °C for 1.5 h. Then, the mixture was subjected to phenol / chloroform (1:1) (Sigma-Aldrich, St. Louis, MO, USA) extraction and ethanol precipitation. Finally, the purified RNA was dissolved in RNase-free water and stored at − 80 °C.

RNA sequencing (RNA-seq) and identification of differentially expressed genes

The total RNA samples from the WT-RS218 and Δprc-RS218 16-h cultures (three independent culture samples for each strain) were subjected to cDNA library construction (paired-end) using a TruSeq RNA sample preparation kit (Illumina, San Diego, CA, USA). Sequencing was performed with the Illumina Solexa platform according to the manufacturer’s protocol (Illumina, San Diego, CA, USA). Qualified reads were aligned to the reference genome of the ExPEC strain UTI89 because the genomes of RS218 and UTI89 are very similar [40, 45]. The transcriptional level of gene expression was calculated and normalized by fragments per kilobase of transcript per million mapped reads (FPKM) [46]. The differentially expressed genes between the wild-type strain and the prc mutant (P value < 0.05) were analyzed by Regulatory Network Interactions of RegulonDB database [47] and classified into different regulons (Table 3).

Table 3 The differentially expressed flagellum- and five ESSSs-related genes between WT-RS218 and Δprc-RS218 by RNA-seq analysis
Full size table

Real-time quantitative PCR (qPCR) and reverse transcription-PCR (RT-PCR)

The purified RNA was reverse transcribed into cDNA by using random hexamer primers and Moloney murine leukemia virus (M-MLV) reverse transcriptase according to the manufacturer’s instructions (Invitrogen, Carlsbad, CA, USA). For qPCR, the cDNA and primers were mixed with KAPA SYBR FAST qPCR Master Mix (Kapa Biosystems, Boston, MA, USA) and then subjected to PCR using a StepOnePlus™ Real-Time PCR System (Applied Biosystems, Carlsbad, CA, USA). The expression levels of the genes were normalized to those of ftsZ. For RT-PCR, the cDNA was subjected to PCR amplification using Taq polymerase for 25 cycles. The resulting products were analyzed by gel electrophoresis and visualized by ethidium bromide (EtBr) staining. The primers used for these assays are shown in Additional file 1: Table S1.

Statistical analysis

Animal UTI experiments were analyzed by using a nonparametric Wilcoxon matched-pair test. The statistical significance of the other experiments were analyzed by unpaired two-tailed Student’s t test. A P value of < 0.05 was considered statistically significant. *, P < 0.05; **, P < 0.01; ***, P < 0.001; and ****, P < 0.0001. (GraphPad Prism 7; GraphPad Software Inc., La Jolla, CA, USA).

Results

Prc deficiency decreases the ability of extraintestinal pathogenic Escherichia coli (ExPEC) to cause urinary tract (UT) infections (UTIs)

We have previously shown that Prc is required for ExPEC to cause a high level of bacteremia [4]. Since ExPEC is the most common cause of UTIs, in the present study, we further investigated whether Prc contributes to ExPEC UTIs. The ability of a prc mutant of the ExPEC strain RS218 (Δprc-RS218) to colonize the UT was compared to that of an otherwise wild-type lacZ mutant (ΔlacZ-RS218). Deletion of lacZ did not affect the ability of ExPEC to colonize the UT (data not shown). After transurethrally infecting a 1:1 mixture of the two bacteria into mice for 48 h, the bacterial loads in the bladder and kidney were determined. Δprc-RS218 showed significantly lower bacterial counts than the otherwise wild-type strain in the bladder and kidney (Fig. 1a). In addition, trans-complementation with the prc gene significantly restored the ability to colonize the organs (Fig. 1b). Similarly, prc mutants of the ExPEC strains CFT073 and UTI89, which were isolated from patients with UTIs, showed decreased abilities to colonize the UT, and trans-complementation of the mutants with the prc gene restored their bacterial counts in the organs (Fig. 1c, d, e, and f). These results demonstrated that Prc contributes to the pathogenesis of UTIs.

Fig. 1
figure 1

Transurethral cochallenge of mice with the ExPEC strains with or with the prc mutation. (a, c, and e) Cochallenges of mice with ExPEC prc mutants and their otherwise wild-type strains. Equal numbers of the prc mutants of the indicated ExPEC strains and their otherwise wild-type strains (the lacZ mutants) were transurethrally inoculated into mice. At 48 h post-inoculation (hpi), bacterial colonization levels of the bladder and kidney were determined. (b, d, and f) Cochallenges of mice with the indicated ExPEC prc mutants and their corresponding complemented strains. lacZ and prc double deletion mutants (∆lacZ∆prc) harboring the prc-encoding plasmid pPrc (Table 1) served as the complemented strains. Equal numbers of the prc mutants harboring the empty vector pCL1920 and the corresponding complemented strains were transurethrally coinoculated into mice. The bacterial counts of each strain in the bladders and kidneys were determined at 48 hpi. The bacterial counts of the mutants and the otherwise wild-type or complemented strains in the same organs were differentiated and enumerated by spreading the homogenized infected tissue onto LB agar plates with IPTG and X-gal. Horizontal bars indicate the median level of the bacterial counts. The dotted line represents the limit of detection

Full size image

Prc deficiency alters the expression of outer membrane (OM)-associated proteins (OMPs) in ExPEC

It has been demonstrated that deletion of prc alters the OMP profile in RS218 [4]. As shown in Fig. 2, prc deletion changed the OMP profiles of CFT073 and UTI89 as well. These findings suggested that alteration of OMP expression is a common outcome of Prc deficiency in ExPEC strains.

Fig. 2
figure 2

The OMP profiles of the ExPEC strains CFT073 and UTI89 and their prc mutants. The outer membrane proteins of the indicated bacteria were separated by 10% SDS-PAGE and then subjected to silver staining

Full size image

The OMPs of pathogenic bacteria are often involved in the host-bacterium interaction during the course of infections [48]. This fact led us to speculate that the altered OMP expression in the prc− ExPEC strains may be responsible for the defect in UT colonization. To determine the OMPs differentially expressed in ExPEC with and without prc, the proteomic profiles of WT-RS218 and Δprc-RS218 were determined by liquid chromatography-tandem mass spectrometry (LC/MS/MS). The levels of 25 and 32 proteins were shown to be significantly downregulated and upregulated, respectively, by the prc deletion (Additional file 2: Table S2). Among them, 2 OMPs (FliC and SlyB) were upregulated in Δprc-RS218, while 6 OMPs (Spr, BamA, Tsx, TolC, Ag43 and TraT) were downregulated in the prc mutant (Table 2).

Prc deficiency decreases the flagellin level and motility of ExPEC

While identified in the LC/MS/MS analysis, FliC (flagellin) is the major component of the flagellar filament, and flagellum-mediated motility is required for ExPEC to cause UTIs [12,13,14,15]. Thus, our study subsequently focused on the FliC protein.

To confirm the LC/MS/MS results regarding FliC, the levels of FliC in WT-RS218 and Δprc-RS218 were further determined by western blot analysis with an anti-FliC antiserum. Consistently, Δprc-RS218 showed lower levels of FliC in total bacterial lysate and the OM fraction than WT-RS218 (Fig. 3a left panel). In addition, the prc mutant showed a significantly lower motility than the wild-type strain (Fig. 3a right panel). Trans-complementation of Δprc-RS218 with the prc gene restored the expression of FliC and motility to the levels of those of WT-RS218. Similar phenotypes were also shown in CFT073 and UTI89 (Fig. 3b and b). These results demonstrate that Prc deficiency reduces flagellin (FliC) expression and suppresses bacterial motility in ExPEC.

Fig. 3
figure 3

The levels of FliC expression and motility of RS218, CFT073, UTI89, and their prc mutants. The effects of prc deletion on FliC expression and motility in the ExPEC strains RS218 (a), CFT073 (b), and UTI89 (c). The results of western blot analyses of the total cell lysates and the OM fractions of the indicated bacterial strains are shown in the left panels of (a), (b), and (c). OmpA served as a loading control. The motilities of the indicated strains on 0.3% agar plates are shown in the right panels of (a), (b), and (c). pCL1920, the empty plasmid vector; pPrc, the plasmid pCL1920 harboring prc (Table 1). The asterisk indicates the FliC protein in the total lysate of CFT073

Full size image

Prc deficiency decreases the expression of the flagellar regulon

fliC is located at the lowest level (class 3) of the transcriptional hierarchy of the flagellar regulon [16]. We investigated whether Prc deficiency affects the expression of this regulatory cascade. Δprc-RS218 exhibited lower expression levels of the class 1 (flhD), class 2 (fliE, fliF, flhA, flgE, flgM, fliM, fliT, and fliA) and class 3 (fliC and motA) genes than WT-RS218 (Fig. 4). Trans-complementation of Δprc-RS218 with prc restored the expression of these genes, suggesting that Prc deficiency results in the downregulation of all classes of genes in the regulon. Given that the top master operon flhDC governs all genes in this regulon, it is likely that prc deletion suppresses flhDC expression to cause the reduced motility. To assess this speculation, we examined the motilities of Δprc-RS218 strains with different levels of flhDC expression. To do so, ∆prc-RS218 was transformed with a flhDC-containing plasmid, pBAD-FlhDC (Table 1). Because the flhDC operon in this plasmid was driven by an arabinose-inducible promoter, arabinose treatment dose-dependently induced the expression of FlhDC (Fig. 5a). The induction of flhDC expression also dose-dependently increased the FliC level and bacterial motility (Fig. 5a and b). These results supported that through suppressing the expression of flhDC, prc deletion downregulates the whole flagellar regulon and consequently decreases bacterial motility.

Fig. 4
figure 4

The expression of the flagellar regulon in the wild-type, prc mutant, and complemented strains of ExPEC. The transcript levels of the class 1 gene flhD, class 2 genes (fliA, flgE, flhA, fliF, fliM, fliE, fliT and flgM) and class 3 genes (fliC, motA, and tar) were determined by qPCR. The transcript levels of the genes in each strain, which were normalized to those of the housekeeping gene ftsZ, are presented as the relative levels compared to those in WT-RS218/pCL1920. The results were derived from experiments performed in triplicate and are shown as the means ± standard deviations. pCL1920, the empty plasmid vector; pPrc, the plasmid pCL1920 harboring prc (Table 1). The asterisks indicate significant differences (P values < 0.05) of the comparisons between WT-RS218/pCL1920 and Δprc-RS218/pCL1920 as well as between Δprc-RS218/pCL1920 and Δprc-RS218/pPrc

Full size image
Fig. 5
figure 5

Effects of increased FlhDC expression on the FliC level and motility in the prc mutant. FliC and HA-tagged FlhD levels (a) and the motility (b) of ∆prc-RS218/pBAD-FlhDC with different levels of L-arabinose treatment. Each motility quantitative result was derived from experiments performed in triplicate and is presented as the means ± standard deviations. The plasmid pBAD-FlhDC harbored HA-tagged flhD and His-tagged flhC genes that were under the control of an arabinose-inducible promoter (Table 1). The levels of FliC, FlhD, and OmpA were determined by western blot analyses with a rabbit anti-FliC antiserum and an anti-HA antibody and mouse anti-OmpA antiserum, respectively. The levels of OmpA served as loading controls

Full size image

The defective motility is responsible for the decreased ability of the prc mutant to cause UTIs

To further investigate whether the defective motility caused by prc deletion contributes to the attenuated ability to cause UTIs, we promoted the motility of prc− ExPEC strains and then assessed their abilities to cause UTIs. The plasmid pFlhDC, which can constitutively overexpress flhDC, was introduced into the prc− strains of RS218, CFT073, and UTI89 and was able to increase their motilities (Fig. 6a). The pFlhDC-harboring prc mutants showed higher levels colonization than the corresponding empty vector-harboring strains (Fig. 6b, c, and d). These results demonstrated that the decreased motility is responsible for the defective ability of prc mutants to colonize the UT.

Fig. 6
figure 6

Effect of motility increase on the ability of the ExPEC prc mutants to cause UTIs. (a) Construction of prc mutants constitutively overexpressing flhDC. The ∆lacZ∆prc strains of the ExPEC strains RS218, CFT073, and UTI89 were transformed with the plasmid pFlhDC (Table 1). The flhDC operon encoded in this plasmid was fused with a lac promoter. Without any induction, the flhDC overexpression driven by leaky promoter activity was strong enough to improve the bacterial motility of the prc mutants. (b, c, and d) Transurethral cochallenge of mice with the prc mutants of the indicated ExPEC strains with and without flhDC overexpression. At 48 hpi, the bacterial counts in the bladders and kidneys were enumerated and differentiated by spreading the homogenized infected tissue onto LB agar plates with IPTG and X-gal. Horizontal bars indicate the median level of the bacterial counts. The dotted line represents the limit of detection

Full size image

Deficiency in Prc increases the expression of the genes governed by some ESSSs

To investigate how Prc deficiency results in the downregulation of flhDC transcription, comparative transcriptome analysis of ∆prc-RS218 and WT-RS218 by RNA-seq was performed. The prc deletion significantly affected the transcriptome of ExPEC, in which 152 and 365 genes were upregulated and downregulated, respectively (Additional file 3: Table S3). Consistently, many genes in the flagellar regulon were shown to be significantly downregulated in the prc mutant (Table 3). In addition, genes known to be able to be upregulated by some extracytoplasmic stress signaling systems (ESSSs), including the σE, RcsCDB, CpxA-CpxR, QseB-QseC, and EnvZ-OmpR systems, showed significantly higher expression levels in ∆prc-RS218 than in WT-RS218 (Table 3). The differential transcript levels of the representative genes governed by these ESSSs were further confirmed by RT-PCR analysis (Fig. 7). These findings suggest that these ESSSs may be activated in the prc mutant. It is likely that the activated ESSSs contribute to the reduced motility.

Fig. 7
figure 7

The expression of the genes positively regulated by the extracytoplasmic signaling systems in WT-RS218 and Δprc-RS218. The transcriptional levels of the genes in the σE (yiiS, hpf, and fkpA), RcsCDB (yjbE and osmC), CpxA-CpxR (aroG and mviM), QseB-QseC (qseB), and EnvZ-OmpR (ompC) regulons [22, 49,50,51,52,53,54,55] were determined by RT-PCR and visualized on agarose gels after EtBr staining. ftsZ served as a housekeeping gene internal control

Full size image

Blocking the activation of the σE or RcsCDB system partially restores the motility of the prc mutant of ExPEC

If the ESSSs play roles in suppressing the motility of the prc mutant, blocking their activation may relieve the suppression. As deletion of the degS, rcsB, cpxR, qseB, and ompR genes blocks the activation of the σE, RcsCDB, CpxA-CpxR, QseB-QseC, and EnvZ-OmpR systems [35, 49, 50, 56, 57], respectively, the deletion of these genes was introduced into Δprc-RS218 to inactivate the corresponding systems. As shown in Fig. 8a, the degS and rcsB deletions partially restored the motility of Δprc-RS218, while the cpxR, qseB, and ompR deletions showed no significant effect on bacterial motility. Complementation of the double mutant strains ΔdegsΔprc-RS218 and ΔrcsBΔprc-RS218 with degS and rcsB in the chromosomal lacZ locus, respectively, decreased the motilities of these strains to the level of that of the prc single mutant (Fig. 8b and c), suggesting that activation of the σE and RcsCDB systems is involved in the decreased motility of the prc mutant. Similar results were also shown in CFT073 (Fig. 8d and e).

Fig. 8
figure 8

The motilities of the prc mutant strains with inactivated extracytoplasmic signaling pathways. (a) The motilities of the prc mutant strains of RS218 with a blocked σE, RcsCDB, CpxA-CpxR, QseB-QseC, or EnvZ-OmpR system. degS, rcsB, cpxR, qseB, and ompR deletions were introduced into Δprc-RS218 to block the σE, RcsCDB, CpxA-CpxR, QseB-QseC, and EnvZ-OmpR signaling pathways, respectively. (b) The motility of the ΔdegsΔprc-RS218 strain complemented with degS in the chromosomal lacZ locus. (c) The motility of the ΔrcsBΔprc-RS218 strain complemented with rcsB in the chromosomal lacZ locus. (d and e) The motilities of the prc mutant strains of CFT073 with an inactivated σE (d) or RcsCDB (e) system

Full size image

The activated σE and RcsCDB systems are involved in the suppression of FliC expression and flhDC transcription in Δprc-RS218

We further investigated whether inactivation of the σE or RcsCDB systems could restore the expression of flhDC and FliC in the prc mutant. yiiS and yjbE are the genes positively regulated by the σE and RcsCDB systems, respectively. Thus, the expression levels of these genes can reflect the activation levels of the corresponding signaling systems [51, 52]. In comparison with Δprc-RS218, the double mutants ΔdegsΔprc-RS218 and ΔrcsBΔprc-RS218 showed lower expression of yiiS and yjbE (Fig. 9a and b). This result indicates that the prc deletion-induced σE and RcsCDB activation was blocked in ΔdegsΔprc-RS218 and ΔrcsBΔprc-RS218, respectively. In addition, in comparison with Δprc-RS218, both ΔdegsΔprc-RS218 and ΔrcsBΔprc-RS218 showed increased levels of flhDC (Fig. 9a and b) and FliC expression (Fig. 9c and d). These results suggest that the activated σE or RcsCDB systems contribute to the suppressed flagellar expression, resulting in the defective motility in the prc mutant.

Fig. 9
figure 9

The effects of σE and RcsCDB inactivation on the transcript levels of the flhDC operon and the FliC levels in the prc mutant. (a and b) The mRNA levels of flhD, yiiS, and yjbE in the indicated strains determined by qPCR. The levels of yiiS and yjbE reflect the activation levels of the σE and RcsCDB systems, respectively. The mRNA level of each gene in a strain was normalized to the ftsZ level and presented as a relative level compared to that in WT-RS218. The results were derived from experiments performed in triplicate and are shown as the means ± standard deviations. (c and d) The levels of FliC in the indicated strains were determined by western blot analysis with a rabbit anti-FliC antiserum. OmpA levels served as loading controls

Full size image

The activated σE system suppresses motility and flhDC transcription in wild-type ExPEC

It has been previously shown that activation of the RcsCDB system can downregulate the expression of flhDC and thus suppress motility in E. coli without prc mutation [21, 24]. This finding suggests that the activation of the RcsCDB signaling pathway alone in the prc mutant strain is sufficient to cause the suppression, requiring no other signals induced by the prc mutation. To the best of our knowledge, this report is the first study demonstrating that the σE system is involved in motility suppression in the prc mutant of E. coli (Fig. 8a, b, and d). However, it remained unclear whether activation of the σE system could suppress bacterial motility in an E. coli strain with an intact Prc. We assessed the motility and flhDC expression of a prc+ RS218 strain (ΔrseA-RS218) in which the σE system was activated by deleting the anti-σE factor RseA [58]. As shown in Fig. 10a and b, activation of the σE system in the prc+ background significantly decreased bacterial motility and suppressed flhDC expression. These findings suggested that without other prc deletion-induced signals, activation of the σE system was sufficient to suppress flhDC expression and thus bacterial motility in the prc mutant.

Fig. 10
figure 10

The effect of σE activation on the motility and transcript level of flhD in the prc+ ExPEC strain. (a) Motility diameters of the indicated strains. (b) Relative mRNA levels of flhD and yiiS determined by qPCR. The yiiS gene served as the reporter of σE activation. The mRNA level of each gene in a strain was normalized to the ftsZ level and presented as a relative level compared to that in WT-RS218 harboring pBAD (WT-RS218/pBAD). Arabinose (0.2%) was used to induce the expression of RseA that was encoded in the pRseA plasmid (Table 1) and driven by the arabinose-inducible promoter on the plasmid. The results were derived from experiments performed in triplicate and are shown as the means ± standard deviations. pRseA, pBAD harboring rseA

Full size image

The σE and RcsCDB systems can work independently to downregulate bacterial motility

We investigated whether the activation of the σE and RcsCDB systems in ExPEC are sequential (upstream and downstream) events. It has been shown that RcsB overexpression triggers the activation of the RcsCDB system [59]. RS218 was transformed with the RcsB-overexpressing plasmid pRcsB (WT-RS218/pRcsB) to create a RcsCDB-activated strain. WT-RS218/pRcsB showed significantly higher yjbE expression than RS218 harboring an empty plasmid vector (WT-RS218/pBAD), while the two strains showed similar levels of yiiS expression (Fig. 11a). This result suggested that activating the RcsCDB system does not trigger the activation of the σE system. On the other hand, the σE-activated strain ΔrseA-RS218 showed significantly higher yiiS expression than WT-RS218, while these two strains showed similar levels of yjbE expression (Fig. 11b). This finding suggested that activation of the σE system does not trigger the activation of the RcsCDB system. Collectively, these results suggested that the activation of these systems is independent of each other in ExPEC.

Fig. 11
figure 11

The activated σE and RcsCDB systems can work independently to suppress bacterial motility. (a) The mRNA levels of the σE-regulated gene yiiS and the RcsCDB-regulated gene yjbE in the RS218 strain with an unactivated or activated RcsCDB system. Arabinose (0.2%) was used to induce the overexpression of RcsB encoded in pRcsB (Table 1). (b) The mRNA levels of yiiS and yjbE in the bacteria with an unactivated or activated σE system. (c) The mRNA levels of yiiS in WT-RS218, Δprc-RS218, and ΔrcsBΔprc-RS218. (d) The mRNA levels of yjbE in WT-RS218, Δprc-RS218, and ΔdegSΔprc-RS218. The mRNA level of each gene, which was determined by qPCR, in a strain was normalized to the ftsZ level and presented as a relative level compared to that in WT-RS218. The results were derived from experiments performed in triplicate and are shown as the means ± standard deviations. pRcsB, pBAD harboring the rcsB gene driven by the arabinose-inducible promoter on the plasmid

Full size image

We further determined whether the two systems are also independently activated by Prc deficiency. ΔrcsBΔprc-RS218, in which the activation of the RcsCDB system was blocked, and Δprc-RS218 showed a similar level of yiiS expression (Fig. 11c), suggesting that the RcsCDB signal does not contribute to the activation of the σE system in the prc mutant. The σE-inactivated strain ΔdegSΔprc-RS218 showed a slightly higher level of yjbE expression than Δprc-RS218 (Fig. 11d), suggesting that σE does not contribute to RcsCDB system activation in the prc mutant. Taken together, these results indicate that Prc deficiency independently induces the activation of the σE and RcsCDB systems.

The protease activity of Prc is required for bacterial motility and FliC expression

Since Prc is a protease, we investigated whether the deficiency of the Prc protease function contributes to the decreased motility and FliC expression in the prc mutant. Prc variants with a S430A or K455A substitution are known to lose catalytic activity but still maintain the original protein structure and substrate-binding ability [60]. Trans-complementation of Δprc-RS218 with the catalytic ability-defective variants failed to restore the motility and FliC level (Fig. 12a and b). These results suggested that deficiency in the Prc protease activity is responsible for the reduced flagellum expression and thus the defective motility.

Fig. 12
figure 12

A lack of Prc protease activity was responsible for the defective motility and decreased FliC expression in the prc mutant. (a) Motility diameters of the strains expressing wild-type Prc, Prc S430A, and Prc K455A. The results were derived from three independent experiments and are shown as the means ± standard deviations. (b) The protein levels of FliC, Prc, and OmpA in the bacteria expressing wild-type Prc, Prc S430A, and Prc K455A. The protein levels were determined by western blot analysis with a rabbit anti-FliC antiserum, rabbit anti-Prc antiserum and mouse anti-OmpA antiserum, respectively. The OmpA level served as a loading control. ΔfliC/pCL1920 served as a nonmobile and non-FliC expression control. pCL1920, the plasmid pCL1920, which served as an empty vector control. pPrc, the plasmid pCL1920 harboring prc; pPrc-S430A, the plasmid pCL1920 harboring a mutated prc expressing Prc S430A; pPrc-K455A, the plasmid pCL1920 harboring a mutated prc expressing Prc K455A

Full size image

The accumulation of Spr is responsible for the defective motility and decreased FliC expression in Δprc-RS218

Given that a protease exerts biological function through mainly proteolytic regulation of its substrates, we reasoned that the dysregulation of Prc substrates due to the deficiency in the protease activity may be responsible for the reduced motility in the prc mutant. The Prc substrate Spr has been shown to accumulate in a prc mutant of the commensal E. coli MG1655 [38]. Similarly, our proteome data showed that Δprc-RS218 expressed a significantly higher level of Spr than WT-RS218 (Table 2). A western blot analysis of Spr in the RS218 strains with or without prc further confirmed this finding (Fig. 13a), suggesting that Spr accumulation is common among the E. coli strains in which Prc is inactivated. We overexpressed Spr to raise the intracellular level of Spr in RS218. Spr overexpression significantly decreased the FliC level (Fig. 13b) and bacterial motility (Fig. 13c), suggesting that Spr accumulation in the prc mutant contributes to reduced motility.

Fig. 13
figure 13

Effects of Spr accumulation on bacterial motility, FliC expression and the activation of the extracytoplasmic signaling systems. (a) The levels of Spr and FliC in the RS218 strains with or without prc. To detect Spr, these strains were modified to express C-terminally 3xFlag-tagged Spr. Western blot analyses with an anti-Flag antibody and a rabbit anti-FliC antiserum were performed to detect the protein levels. The OmpA levels served as loading controls, which were probed with a mouse anti-OmpA antiserum. Spr-3xFlag-RS218, the RS218 strain with the native chromosomal spr fused with a sequence encoding a 3xFLAG tag at the 3′ end; Spr-3xFlag-Δprc-RS218, the Δprc-RS218 strain with the chromosomal spr fused with a sequence encoding a 3xFLAG tag at the 3′ end. (b) The FliC and Spr levels in the RS218 strains with or without the overexpression of the recombinant Spr, which was C-terminally fused with a Flag tag. The Spr protein was detected with an anti-Flag antibody. Arabinose (0.2%) was used to trigger the expression of the recombinant Spr that was encoded in pBAD and driven by the arabinose-inducible promoter in the plasmid. pSpr, pBAD harboring spr fused with a sequence encoding a Flag tag at the 3′ end. (c) Motility diameters of the strains overexpressing Spr. (d) The relative mRNA levels of yiiS and yjbE compared to those in WT-RS218/pBAD. (e) The relative yibE levels compared to those in WT-RS218. (f) The FliC level in WT-RS218, Δprc-RS218, and ΔsprΔprc-RS218. (g) Motility diameters of WT-RS218, Δprc-RS218, and ΔsprΔprc-RS218

Full size image

To determine whether Spr accumulation in ExPEC triggers activation of the RcsCDB and σE systems, the transcription levels of the RcsCDB-regulated yjbE and σE-regulated yiiS genes in an Spr-overexpressing strain (WT-RS218/pSpr) and a strain with normal Spr expression (WT-RS218/pBAD) were determined (Fig. 13d). Overexpression of Spr upregulated yjbE, but yiiS was not affected. This suggested that Spr accumulation triggers activation of the RcsCDB system but not the σE system. Consistent with this finding, blocking the activation of RcsCDB signaling by deleting rcsB significantly increased the motility of the Spr-overexpressing strain, while blocking σE signaling by deleting degS did not affect bacterial motility (Fig. 13c). Additionally, deletion of rcsB increased the FliC level in the Spr-overexpressing strain (Fig. 13b). These results suggest that Spr accumulation in the prc mutant triggers activation of the RcsCDB system and thus suppresses flagellin expression and bacterial motility.

In addition, we further investigated whether blocking Spr accumulation in the prc mutant affects the activation of the RcsCDB system and motility. We blocked Spr accumulation in the prc mutant by deleting the spr gene (ΔsprΔprc-RS218). Similar to wild-type RS218, ΔsprΔprc-RS218 showed significantly lower yibE expression than Δprc-RS218 (Fig. 13e), further supporting that Spr accumulation results in the activation of the RcsCDB system in the prc mutant. However, the FliC level (Fig. 13f) and motility (Fig. 13g) of ΔsprΔprc-RS218 were not significantly higher than those of Δprc-RS218. These results suggest that deleting spr in the prc mutant may cause other pleotropic effects that are able to downregulate FliC expression and motility, which can offset the effects of the downregulated RcsCDB activation.

Discussion

This study reveals for the first time that the periplasmic protease Prc in ExPEC contributes to the pathogenesis of UTIs through maintaining intact bacterial motility, which is required for ExPEC to colonize the bladder and kidney [12,13,14,15]. In addition, the mechanism of how Prc deficiency interferes with the motility in ExPEC has been elucidated (see the model in Fig. 14). Lack of Prc in the bacteria triggers σE and RcsCDB signaling, which in turn negatively regulate the expression of the master operon flhDC of the flagellar regulon, leading to decreased flagellum expression and hindered bacterial motility. To our knowledge, this report is also the first study demonstrating that σE signaling negatively regulates the expression of flagella, while activated RcsCDB signaling has previously been shown to be able to suppress this bacterial structure [21]. The accumulation of Spr, a Prc substrate, is responsible for the activated RcsCDB signaling in the ExPEC prc mutant. These results demonstrate that Prc-mediated proteolytic regulation of the intracellular substrate is critical for sufficient ExPEC motility to cause UTIs.

Fig. 14
figure 14

The model describing how deficiency in the Prc protease leads to defective bacterial motility in E. coli. Prc protease deficiency interferes with proteolytic regulation of its substrates in the bacterial envelope, leading to uncontrolled levels of Prc substrates, such as Spr. The accumulation of the substrates may alter the OMP profile, leading to compromised OM integrity and disturbing peptidoglycan biogenesis. The resulting disturbance in the envelope triggers activation of the σE or RcsCDB extracytoplasmic stress response systems. The activated σE and RcsCDB systems decrease flagellar biosynthesis and thus bacterial motility

Full size image

The RcsCDB system monitors damage of the OM and peptidoglycan layer [24, 61,62,63,64]. It is highly likely that the uncontrolled Spr level in the prc mutant activates the RcsBCD system through altering the peptidoglycan structure and interfering with OM integrity. This notion is consistent with the known physical function of Spr and the phenotype resulting from the overexpression of Spr in E. coli [65, 66]. Spr is an OM protein [67] and a peptidoglycan hydrolase involved in maintaining the stability of the bacterial peptidoglycan structure [65]. It has been shown that overexpression of Spr in E. coli interferes with peptidoglycan biogenesis, leading to decreased peptidoglycan crosslinkage with the OM, which consequently destabilizes the bacterial envelope [66]. In addition, the high level of Spr located on the OM may be a contributing factor to the compromised OM integrity in the prc mutant of ExPEC.

Dual molecular signals are required to activate the σE signaling system: the periplasmic accumulation of denatured OMPs and LPS [36, 68]. Thus, the activation of this signaling system in the prc mutant suggests that Prc deficiency causes the accumulation of denatured OMPs and LPS in the periplasmic space. In agreement with this hypothesis, we found that overexpression of DegQ, which is a periplasmic chaperone able to renature misfolded proteins in the periplasm [69], could downregulate the σE activation and upregulate the FliC expression and motility of the E. coli prc mutant (Additional file 4: Figure S1). In addition, construction of the OM requires the transport of OMPs and LPS through the periplasmic space into the OM. In the prc mutant, the periplasmic accumulation of these components suggests that Prc deficiency may hinder their transport from the periplasm to the OM. The global proteomic analysis showed that the prc mutant (Δprc-RS218) exhibited a significantly higher level of BamA expression than the wild-type strain (Table 2). BamA is the essential component of the β-barrel protein assembly machinery (BAM complex), which is responsible for the assembly of OMPs and the LPS transporter proteins in the OM [70, 71]. The increased BamA expression may be a compensatory response for the mutant since the complex is required for the transport of the OMPs and LPS transporters from the periplasm to the OM.

In the prc mutant, both the RcsCDB and σE systems suppressed bacterial motility through downregulating the expression of flhDC. It has been known that RcsB can directly bind to the promoter of flhDC to suppress its transcription when the RcsCDB system is activated [21]. However, how σE signaling downregulates this operon remains to be elucidated. The RNA polymerase holoenzyme (holo-RNAP) is composed of an σ subunit and a core RNA polymerase (co-RNAP), which are responsible for promoter selectivity and RNA synthesis, respectively. In bacterial cells, different types of σ factors compete for a limited pool of common co-RNAP to transcribe a set of genes driven by their cognate promoters [72]. Transcription of the flhDC operon is dependent on the housekeeping sigma factor σ70 [73]. It is likely that in the prc mutant, the activated σE system allows the σE factor to compete with σ70 for co-RNAP, thus resulting in the downregulation of the flhDC operon. Alternatively, σE may trigger the expression of unknown bacterial factors capable of downregulating flhDC transcription.

The present study demonstrates that constitutive activation of σE and RcsCDB signaling attenuates ExPEC in UTIs through suppressing flagellum-mediated motility. However, it has been shown that blocking σE signaling will also reduce the bacterium’s ability to cause UTIs [74]. These findings suggest that a tunable σE signaling system rather than a constantly activated or silenced one is required for the whole virulence of the pathogen during infections. To cause UTIs, ExPEC need to accomplish multiple pathogenic steps, such as adhering to the epithelium lining of the UT, disseminating within the UT, and evading the host immune responses, etc. [75]. The existence of a certain bacterial factor may benefit one pathogenic step but hinder another. Strict regulation of the expression of such factors may also be required for the whole virulence of the pathogen. While flagella are necessary for E. coli to disseminate within the UT during infections, flagellin, the major component of flagella, is a potent immunogen able to activate the immune response via TLR5, which may cause the clearance of the invading pathogens. Therefore, σE signaling may need to remain tunable so that it can work in coordination with flagellar expression to achieve a successful UTI. In addition, inactivation of the RcsCDB system has been shown to decrease the ability of ExPEC resistance to serum-mediated killing [76]. This finding suggest that a tunable RcsCDB system may also be essential for the full virulence of ExPEC. Thus, hindering the proper function of the σE and RcsCDB signaling systems would be a potential strategy to fight against the bacterial infections.

The decreased motility may not be fully responsible for the defective ability of the ExPEC prc mutant to cause UTIs. Our previous study demonstrated that prc deletion decreases the resistance of ExPEC to complement-mediated killing in the bloodstream [4]. It has been suggested that pathogenic E. coli could be opsonized by the complement system in the UT [77]. Since bacterial opsonization could facilitate phagocytosis by phagocytes, leading to the elimination of invading bacteria in the tissue, defective resistance to the complement system may also attenuate the prc mutant in UTIs. In agreement with this hypothesis, we found that increasing the motility of Δprc-RS218 by overexpressing flhDC to the level of flhDC expression in WT-RS218 could not restore the mutant’s UT colonization to the level of the wild-type strain (data not shown). On the other hand, the decreased motility was not responsible for the prc mutant’s defect in the resistance to killing mediated by the complement system because increasing the motility of Δprc-RS218 did not restore bacterial resistance to complement-mediated serum killing (data not shown).

In addition to FliC, several other outer membrane proteins whose expression levels were significantly affected by the prc deletion were found in the LC/MS/MS analysis (Table 2). The altered expression levels of the proteins may also contribute the attenuated virulence of the ExPEC prc mutant. For example, SlyB was shown to be downregulated in the prc mutant of RS218 (Table 2). It is known that deletion of SlyB in Burkholderia multivorans attenuates bacterial iron uptake ability and compromises the OM integrity [78]. Given that the iron uptake ability and an intact OM integrity are important for pathogenic bacteria to invade hosts [79,80,81], it is worth further investigating whether the downregulated SlyB expression in the ExPEC prc mutant decreased the bacterial ability to cause UTIs. In addition, Tsx was shown to be upregulated in the prc mutant of RS218. The outer membrane porin Tsx is essential for E. coli to uptake of a gyrase inhibitor antibiotic, albicidin [82,83,84]. It has been known that deletion of prc increases the susceptibilities of E. coli to multiple antibiotics [85]. The increased Tsx level may be one of the contributing factors.

Conclusions

In addition to our previous finding that lacking of Prc resulted in decreased ability of ExPEC to cause bacteremia [4], we demonstrated herein that such defect can also diminish the ability of ExPEC to cause UTIs. Given prc contributes to ExPEC infections in different extraintestinal host tissues as found in the bloodstream and UT, our study strengthens an idea that Prc or Prc homologs may be a potential antimicrobial target for developing a novel strategy in managing ExPEC or other bacterial infections. In addition, the σE and RcsCDB systems, which are responsible for the defect of the prc mutant in causing UTIs, are potential antimicrobial targets in the same light.