EsiB, a Novel Pathogenic Escherichia coli Secretory Immunoglobulin A-Binding Protein Impairing Neutrophil Activation

ABSTRACT

In this study, we have characterized the functional properties of a novel Escherichia coli antigen named EsiB (E. coli secretory immunoglobulin A-binding protein), recently reported to protect mice from sepsis. Gene distribution analysis of a panel of 267 strains representative of different E. coli pathotypes revealed that esiB is preferentially associated with extraintestinal strains, while the gene is rarely found in either intestinal or nonpathogenic strains. These findings were supported by the presence of anti-EsiB antibodies in the sera of patients affected by urinary tract infections (UTIs). By solving its crystal structure, we observed that EsiB adopts a superhelical fold composed of Sel1-like repeats (SLRs), a feature often associated with bacterial proteins possessing immunomodulatory functions. Indeed, we found that EsiB interacts with secretory immunoglobulin A (SIgA) through a specific motif identified by an immunocapturing approach. Functional assays showed that EsiB binding to SIgA is likely to interfere with productive FcαRI signaling, by inhibiting both SIgA-induced neutrophil chemotaxis and respiratory burst. Indeed, EsiB hampers SIgA-mediated signaling events by reducing the phosphorylation status of key signal-transducer cytosolic proteins, including mitogen-activated kinases. We propose that the interference with such immune events could contribute to the capacity of the bacterium to avoid clearance by neutrophils, as well as reducing the recruitment of immune cells to the infection site.

IMPORTANCE

Pathogenic Escherichia coli infections have recently been exacerbated by increasing antibiotic resistance and the number of recurrent contagions. Attempts to develop preventive strategies against E. coli have not been successful, mainly due to the large antigenic and genetic variability of virulence factors, but also due to the complexity of the mechanisms used by the pathogen to evade the immune system. In this work, we elucidated the function of a recently discovered protective antigen, named EsiB, and described its capacity to interact with secretory immunoglobulin A (SIgA) and impair effector functions. This work unravels a novel strategy used by E. coli to subvert the host immune response and avoid neutrophil-dependent clearance.

Introduction

Escherichia coli is able to colonize the human gastrointestinal tract soon after birth, often persisting for decades and providing mutual benefit to the host and bacterium. Although E. coli strains are usually regarded as commensals, certain isolates have acquired specific virulence traits that confer the ability to colonize new niches and to cause a wide variety of diseases (1). Pathogenic E. coli strains can be classified as either intestinal (InPEC) or extraintestinal (ExPEC), depending on the site of infection (2). The ability of E. coli to cause a wide range of diseases is partly due to its variable genome and its ability to gain or lose virulence attributes by horizontal gene transfer.

Host defense against pathogenic E. coli colonization of mucosal surfaces depends at least in part on the role of secretory immunoglobulin A (SIgA) and other opsonins in preventing bacterial adherence. SIgA is a hydrophilic, negatively charged molecule due to the predominance of hydrophilic amino acids in its Fc region and to the abundant glycosylation of both IgA and its secretory component. Therefore, SIgA can surround microorganisms with a hydrophilic shell that is repelled by the mucin glycocalyx at mucosal surfaces. SIgA is also able to agglutinate microbes and interfere with bacterial motility by interacting with flagella. In addition, SIgA interacts with bacterial products such as enzymes and toxins, neutralizing their action (3). Although IgA has so far been considered an anti-inflammatory antibody, recent data support its capacity to trigger a plethora of inflammatory functions by interacting with the myeloid IgA Fc receptor FcαRI (CD89), which is constitutively expressed on cells of the myeloid lineage, including neutrophils, monocytes, and macrophages (4). Although mucosal immunity is crucial to the control of E. coli infections, no IgA-binding proteins (IgA-BPs) have been so far reported for any of the pathotypes.

Recently, a comparative genome analysis of ExPEC and nonpathogenic E. coli strains, named subtractive reverse vaccinology, led to the identification of nine potential vaccine candidates able to confer protection in a murine model of sepsis (5). Although some of the protective antigens have already been functionally described in the literature (6, 7), most of them have been assigned only putative or hypothetical functions; therefore, their experimental characterization could contribute to a greater understanding of ExPEC pathogenesis.

In this study, we present data on the functional characterization of the protective antigen c5321 (5) as a novel E. coli Sel1-like repeat (SRL) protein able to bind SIgA and inhibit its effector functions. This antigen has been named EsiB (E. coli secretory immunoglobulin A-binding protein), and it potentially contributes to ExPEC evasion of the host immune system.

RESULTS

The esiB gene is preferentially associated with ExPEC isolates.

The presence of the c5321 gene, here named esiB, was evaluated by PCR and bioinformatics analyses in a collection of 267 strains representative of different E. coli pathotypes, including 119 ExPEC, 117 InPEC, and 31 nonpathogenic strains. This panel comprised both isolates and complete genome sequences available in the NCBI database of human (n = 190) and animal (n = 77) origin (see Table S1 in the supplemental material). This analysis revealed that the gene was preferentially associated with ExPEC strains (29/119, 24.3%), while it was found in only a very low number of InPEC (6/117, 5.1%) and nonpathogenic (1/31, 3.2%) strains (Table 1). Moreover, multilocus sequence typing (MLST) and ancestral group analysis of the strains PCR positive for EsiB revealed that the gene was almost exclusively related to B2 strains, while its presence was not associated with a specific sequence type (see Fig. S1 and Table S1).

TABLE 1 .

Distribution of esiB gene among pathogenic and nonpathogenic strains

Pathotype	No. positive	Total no.	Prevalence, %
ExPEC	29	119	24.3
InPEC	6	117	5.1
Nonpathogenic	1	31	3.2
All	36	267	13.4

Genome sequence analysis of the esiB locus indicated that the gene was flanked by genes of unknown or putative functions (Fig. 1A). When absent, the ChpBS toxin-antitoxin system was found in the same locus (Fig. 1B), but in several strains, both esiB and chpBS genes were absent (Fig. 1C). Of interest, the esiB gene was edged by short direct repeats with a partial inverted symmetry (Fig. 1D), suggesting that a homologous recombination event might have occurred. In the complete genomes analyzed where esiB was absent, only one copy of the repeat was found. The analysis of 47 complete E. coli genomes present in the NCBI databases revealed that 32% of strains lacked both esiB and chpBS genes, while 53% contained the chpBS gene and 15% contained the esiB gene.

FIG 1 .

Genome sequence analysis of esiB locus. (A to C) In extraintestinal pathogenic E. coli (ExPEC) strain CFT073, the esiB gene (red arrow) is flanked by genome regions composed of conserved genes (blue arrows). The esiB upstream genes (ytfM, ytfN, and ytfP) are annotated as hypothetical proteins, while genes downstream of esiB (ppa and ytfQ to ytfF) are coding for the membrane component of a putative sugar ABC transporter system (A). When esiB is absent, in the K-12 nonpathogenic strain MG1655, the ChpBS toxin-antitoxin system (yellow arrows) can be found in the same context (B), but in several intestinal pathogenic (InPEC) strains, both esiB and chpBS genes are absent (C). (D) Graphical representation of the esiB gene region flanked by short direct repeats.

EsiB is expressed in vivo during UTIs.

To evaluate the immunogenicity of EsiB in vivo and to confirm its expression during key events of E. coli pathogenesis, we collected a panel of 22 sera from patients suffering from urinary tract infections (UTIs) and measured IgG titers against the antigen by enzyme-linked immunosorbent assay (ELISA). As shown in Fig. 2A, serum IgG titers in UTI patients were significantly higher than those of healthy human donors, suggesting that EsiB is immunogenic and expressed in vivo during E. coli infections. Moreover, immunoblot analysis revealed that recombinant EsiB was specifically recognized by a mouse antiserum raised by immunizing animals subcutaneously with live bacteria, indicating that the antigen is immunogenic during in vivo challenge (see Fig. S2 in the supplemental material). To define the kinetics of EsiB expression under in vitro conditions, we evaluated the mRNA levels in the uropathogenic strain CFT073 grown in Luria-Bertani (LB) medium by real-time quantitative PCR, and we observed that antigen expression was higher in logarithmic phase than in stationary phase (Fig. 2B). A time course analysis of bacteria grown in minimal medium revealed that the level of EsiB transcript remained constant up to 3 h of incubation (exponential phase) and declined during stationary phase (from 4 to 7 h) (see Fig. S3). Although esiB transcription was confirmed in vitro, the protein was not detected under any culture condition tested (data not shown), suggesting that in vitro mRNA levels in CFT073 might be insufficient to produce a detectable amount of the protein. However, confocal immunofluorescence microscopy of paraffin-embedded bladder sections derived from mice transurethrally infected with CFT073 revealed that EsiB is specifically expressed by bacteria associated with the bladder tissue (Fig. 3B). No signal was detected by using a serum raised by immunizing mice with adjuvant alone (Fig. 3A). Of interest, the staining was not evident in all bacteria and anti-EsiB antibodies appeared to recognize the protein at bacterial poles, suggesting a specific localization (Fig. 3C). Imaging data are representative of a qualitative imaging analysis of 10 bladder sections. In addition, by engineering an E. coli BL21(DE3) strain for the overexpression of EsiB under the control of an inducible promoter, we confirmed by both flow cytometry and confocal microscopy analysis (Fig. 4) that the antigen was exposed on the bacterial surface, suggesting that specific expression regulatory events may occur during in vivo infection. Moreover, when we performed Western blotting of the secreted faction from the BL21::esiB strain, we observed no presence of the protein (data not shown), suggesting that EsiB exerts its activity on the bacterial surface, although it cannot be excluded that, during in vivo infection, the protein may be released by bacteria undergoing stress conditions.

FIG 2 .

EsiB is recognized by human sera. (A) Immunoreactivity of human sera from UTI patients toward recombinant EsiB. ELISA was performed by applying the recombinant antigen as a coating on microtiter plates and overlaying wells with specific human sera. IgG binding was revealed by alkaline phosphatase-conjugated goat anti-human IgG. Sera from healthy human volunteers were used as IgG control levels. The data represent the means ± standard deviations of 22 human sera. Titers are reported as serum dilutions. *, comparison of EsiB IgG titers between healthy human volunteers and UTI patients: P ≤ 0.05 (paired Student’s t test). (B) EsiB expression under in vitro conditions. EsiB expression in UPEC strain CFT073 (wild type [wt]) grown in LB medium was evaluated by real-time quantitative PCR, using the isogenic mutant strain CFT073ΔesiB (ko) as a negative control. For relative quantification of gene expression, the starting mRNA copy number of the unknown samples was determined using the comparative ΔΔC_T method, and levels of the different transcripts were normalized to 16S rRNA, used as a housekeeping gene. The graph, representing a typical experiment out of two performed with similar results, shows that EsiB expression is higher in logarithmic phase than in stationary phase.

FIG 3 .

EsiB is expressed on the E. coli surface in vivo. Confocal imaging analysis of EsiB expression in bladder from mice infected with the CFT073 strain. Tissue sections embedded in paraffin were stained with WGA conjugated with Alexa Fluor 647 (violet), while the CFT073 strain was detected using a polyclonal serum raised against inactivated whole bacteria (green). An antiserum raised against the adjuvant alone was used as a negative control (A), while EsiB was stained with a specific rabbit antiserum (B) (red signal). The images are representative of multiple observations, from at least 10 sections. (C) Magnification of EsiB expression in bacteria colonizing the bladder. Staining of bacteria was performed as described for panel A.

FIG 4 .

Constitutive expression of EsiB in strain BL21::esiB postulates the exposure of the protein on the bacterial surface. (A) EsiB expression in the BL21::esiB strain was evaluated by flow cytometry, using the E. coli BL21(DE3) strain as a negative control. The histogram represents a typical experiment out of two performed with similar results. RFI, relative fluorescence intensity. (B) Confocal microscopy of EsiB surface localization in BL21 wild-type (top) and BL21::esiB (bottom) strains. Fixed bacteria were incubated with an anti-EsiB serum and then with the fluorescent secondary antibodies.

EsiB binds to human SIgA by a specific motif.

By solving the crystal structure of EsiB (Protein Data Bank [PDB] entry 4BWR) (D. Urosev, M. Ferrer-Navarro, I. Pastorello, E. Cartocci, L. Costenaro, D. Zhulenkovs, J. D. Maréchal, A. Leonchiks, D. Reverter, L. Serino, M. Soriani, X. Daura, submitted for publication), we observed that the antigen is composed of 11 SLRs (residues 16 to 408), an N-terminal and two C-terminal capping helices, and a (partly helical) C-terminal tail. Considering that SLR proteins have been reported to be involved in immunomodulatory functions and immune evasion strategies (8) and that SIgA has been shown to play a major role in the host response to UTIs (9), we tested the capacity of recombinant EsiB to interact with human SIgA. ELISAs showed that EsiB binds to SIgA through a dose-dependent mechanism, with an estimated dissociation constant (K_d) of approximately 10⁻⁷ M (Fig. 5A). Binding of EsiB to SIgA was confirmed by surface plasmon resonance (SPR) experiments. In brief, SIgA antibodies were covalently immobilized on a CM5 sensor chip via amine coupling and the interaction of increasing concentrations of the recombinant EsiB protein was monitored. The affinity between EsiB and SIgA was calculated by fitting the resulting curves with a 1:1 binding model, revealing an estimated dissociation constant (K_d) of 3 × 10⁻⁷ M (Fig. 5B).

FIG 5 .

EsiB is able to bind to human SIgA. (A) ELISA plates were coated with SIgA and incubated with a dilution series of EsiB. Binding of EsiB was detected using mouse anti-EsiB monoclonal antibodies and peroxidase-conjugated rabbit anti-mouse IgG. ELISA data show that EsiB is able to bind human SIgA with an estimated dissociation constant of approximately 10⁻⁷ M. SIgA binding to an unrelated E. coli antigen (ECOK1_0290) was used as a negative control. The graph represents a typical experiment out of three performed with similar results. (B) Solution-phase EsiB-SIgA binding measured by surface plasmon resonance (BIAcore). Representative BIAcore sensorgrams show the response over time (resonance units [RU]) during the binding of purified recombinant EsiB to immobilized SIgA antibodies. An overlay of three repeated titration series is shown. Fitting by a 1:1 Langmuir model of the collected sensorgrams gave the following averaged kinetic constants: k_on, 1.39 × 10³ M⁻¹ s⁻¹; k_off, 4.2 × 10⁻² 1/s; K_d, 3.02 × 10⁻⁵ M.

To map the IgA-binding site on EsiB, we used an immunocapturing approach that employed superparamagnetic beads (Dynabeads). SIgA was coupled to the surface of the beads, which were then used to capture the EsiB peptide derived from partial digestion of the antigen with trypsin or with GluC in ammonium bicarbonate buffer. Following tryptic digestion, SIgA captured a single peptide with an m/z of 1,881.968 (Fig. 6), which by tandem mass spectrometry (MS/MS) analysis (data not shown) was identified as the fragment 244-VLFSQSAEQGNSIAQFR-260. Digestion with GluC in ammonium bicarbonate buffer resulted in no peptides being captured. In bicarbonate buffer, GluC cuts only after E residues. Therefore, this result suggests that residue E251 and its immediate neighbors might play an important role in the interaction between SIgA and EsiB. Importantly, the peptide relative to the SIgA-binding site is conserved among sequenced pathogenic E. coli strains.

FIG 6 .

Identification of the EsiB IgA-binding motif. (A) Mass spectra obtained during the immunocapturing procedure. The upper spectrum shows the peptide mixture obtained after EsiB digestion with trypsin, and the lower spectrum shows the captured peptide with an m/z of 1,881.968. This peptide corresponds to the segment 244-VLFSQSAEQGNSIAQFR-260. (B) The position of the identified peptide is shown in red on the EsiB structure (PDB entry 4BWR) (Urosev et al., submitted).

EsiB inhibits the SIgA effector functions mediated by neutrophil receptor FcαRI.

SIgA, which is the predominant antibody present in the secretions from mucosal surfaces of the genitourinary tract (10), has been shown to play a major role, together with polymorphonuclear neutrophils (PMNs), in the host response to UTIs (9). Indeed, SIgA can initiate numerous immune effector functions via the neutrophil IgA receptor FcαRI, leading to a plethora of potent eradication mechanisms, such as the production of reactive oxygen species (ROS) and the release of cytokines, chemotactic factors, and inflammatory lipid mediators (11). Therefore, in order to study the biological significance of EsiB binding to SIgA, we investigated the antigen capability to interfere with antibody effector functions.

As SIgA can induce superoxide generation by PMNs in vitro (3), we first assessed the ability of EsiB to hamper the SIgA-mediated triggering of a respiratory burst in neutrophils, characterized by NADPH oxidase activation and degranulation, which can be measured in a chemiluminescence assay (12). Interestingly, we observed that EsiB is capable of blocking the SIgA-triggered oxidative burst at concentrations greater than 1 µg/ml, in a dose-dependent manner (Fig. 7A). In particular, a concentration of 20 µg/ml was sufficient to cause approximately 80% inhibition of ROS production after 30 min of incubation at 37°C (Fig. 7B). Moreover, EsiB did not inhibit a phorbol 12-myristate 13-acetate (PMA)-induced respiratory burst (data not shown), suggesting that its inhibitory effect is specific for SIgA stimulation. Together, these data indicate that EsiB can block the SIgA-triggered oxidative burst, which may be advantageous to the pathogen.

FIG 7 .

EsiB inhibits SIgA effector functions. (A) EsiB inhibits the SIgA-mediated neutrophil oxidative burst in a dose-dependent manner. ROS production by PMNs was measured by chemiluminescence every 30 s for 60 min at 37°C and quantified in arbitrary units (AU). The graph represents a typical experiment out of five performed with similar results. (B) ROS production at 30 min is shown as relative area under the curve (AUC), in the presence of 20 µg/ml of EsiB. (C) EsiB inhibits the SIgA-induced neutrophil chemotaxis. To measure neutrophil chemotaxis, bottom chambers of Transwell supports were filled with supernatants deriving from PMNs incubated with SIgA or EsiB-treated SIgA. Neutrophils were added to the upper chambers. After 1 h at 37°C, cells that had migrated toward the lower compartments were quantified by flow cytometry. The graph represents a typical experiment out of three performed with similar results. ***, P ≤ 0.001; **, P ≤ 0.01. Error bars, SD.

Considering that cross-linking of FcαRI on neutrophils also leads to the release of different chemotactic factors (13), to further assess the physiological relevance of EsiB interaction with SIgA, we tested its ability to interfere with neutrophil migration. Using supernatants deriving from PMN incubation with SIgA or EsiB-treated SIgA as a source of chemotactic stimuli, we quantified neutrophil migration by flow cytometry and found that EsiB, used at a concentration of 30 µg/ml, is able to cause approximately 40% inhibition of chemotaxis (Fig. 7C). These data suggest that EsiB can induce a decreased recruitment of neutrophils at the infection site, which results in avoiding clearance.

EsiB impairs FcαRI-mediated neutrophil activation.

IgA-BPs usually allow bacteria to evade IgA-driven eradication processes by blocking the interaction of IgA with FcαRI. Therefore, to elucidate the mechanism of action of EsiB, we investigated the antigen capacity to block this interaction. SIgA was preincubated with EsiB before addition to neutrophils, and cells were then stained with a fluorescein isothiocyanate (FITC)-labeled anti-human IgA antibody. Flow cytometric analysis revealed that EsiB is able to bind SIgA and that this interaction does not prevent SIgA binding to neutrophils (data not shown), suggesting that EsiB most likely binds to a region on SIgA not overlapping with the FcαRI binding site. EsiB used as a negative control did not bind to cells (data not shown), showing that the inhibition of SIgA-mediated eradication processes was not due to a direct effect of the protein on neutrophils. Together, these data indicate that, unlike most bacterial IgA-BPs, EsiB does not inhibit SIgA-FcαRI complex formation.

Since SIgA binding to PMNs was not hampered by EsiB, we hypothesized that the inhibition of antibody effector functions could be due to EsiB preventing productive FcαRI cross-linking, which is crucial for initiation of signal transduction pathways that lead to neutrophil activation.

To understand whether EsiB could interfere with FcαRI signaling, we analyzed the general tyrosine phosphoprotein pattern in neutrophils stimulated with SIgA alone or in combination with EsiB, and we observed that preincubation with EsiB results in a dose-dependent decrease in the phosphorylation level of different cytoplasmic proteins (Fig. 8A), including phospholipase C-γ (PLC-γ), which has been reported to participate in signal transduction pathways involved in activation of NADPH oxidase and chemotaxis (14). Neutrophil stimulation with EsiB alone did not result in cell activation (data not shown), consistent with the inability of the antigen to bind to PMNs.

FIG 8 .

EsiB interferes with FcαRI signaling. (A) Neutrophil activation was performed by incubating cells with SIgA or EsiB-treated SIgA for 1 min at 37°C. Cells were lysed in 1% Triton X-100 in 20 mM Tris-HCl, pH 8.0, 150 mM NaCl, in the presence of a protease inhibitor cocktail. Equal amounts of proteins from each sample were resolved by 10% SDS-PAGE and transferred to 0.45-µm nitrocellulose filters. Immunoblot assays were carried out using antiphosphotyrosine antibody and peroxidase-conjugated rabbit anti-mouse antibody. Blots were reprobed with loading control antibody after stripping (antiactin). The positions of molecular mass markers (kDa) are indicated on the left. The interaction of EsiB with SIgA resulted in dose-dependent modifications in the tyrosine phosphorylation level of specific cytosolic proteins, including PLC-γ and others not yet identified (*). PLC-γ identity was confirmed by reprobing the blot with anti-PLC-γ1 antibody after stripping (data not shown). The blot represents a typical experiment out of three performed with similar results. (B and C) Immunoblot analysis, using a phosphospecific antibody, of ERK1/2 (B) and p38 (C) phosphorylation in neutrophils stimulated with SIgA or EsiB-treated SIgA. Filters were stripped and reprobed with control anti-ERK (B) and anti-p38 (C) antibodies. The results of a representative experiment are shown. (D) Quantification by laser densitometry of the relative levels of ERK1/2 phosphorylation (the ERK1/2 phosphorylation in control cells was defined as 100%) and p38 phosphorylation (the p38 phosphorylation in control cells was defined as 100%). EsiB affected to a significant extent the basal level of both ERK1/2 and p38 phosphorylation. ***, P ≤ 0.001; **, P ≤ 0.01. Error bars, SD.

To further investigate whether EsiB could selectively affect downstream components of FcαRI signaling, we focused on two members of the mitogen-activated protein kinase (MAPK) family specifically implicated in neutrophil responses: the extracellular signal-regulated kinases 1 and 2 (ERK1/2), involved in cell survival and proliferation, and the stress-activated kinase p38, participating in respiratory burst activity and chemotaxis (15). The analysis of MAPK activation using phosphospecific antibodies showed that EsiB is able to impair, in a dose-dependent manner, FcαRI-dependent phosphorylation of ERK1/2 and p38 (Fig. 8B to D). Collectively, these data suggest that EsiB can interfere with a productive FcαRI signaling, resulting in the inhibition of SIgA effector functions.

DISCUSSION

In this study, we functionally characterized EsiB, an E. coli antigen (previously annotated as c5321) proposed as a candidate for an ExPEC vaccine (5). The recent availability of structural data (Urosev et al., submitted) has raised new perspectives on the possible involvement of EsiB in E. coli pathogenesis and evasion of host immune responses. This hypothesis was in part supported by the analysis of human sera, which showed that patients suffering from UTIs possessed IgG titers against the antigen significantly higher than titers in healthy individuals, suggesting that EsiB triggers an in vivo immune response during E. coli infection of the urinary tract. This is in agreement with the fact that the esiB gene was prevalently associated with ExPEC strains (B2 group), with uropathogenic pathotypes being the most well represented. The fact that EsiB was expressed by bacteria in a detectable amount only under in vivo infection conditions suggests that the antigen may have a role during a specific stage of pathogenesis, although the regulatory mechanism remains unknown. The antibacterial defense of the urinary tract depends on two key players: neutrophils, which are crucial effectors of innate immunity, and SIgA, which not only prevents pathogen adherence to mucosal surfaces but also triggers a wide variety of elimination mechanisms elicited upon cross-linking of FcαRI on neutrophils, including the activation of a respiratory burst and the release of cytokines, chemotactic factors, and inflammatory lipid mediators (9, 11). The evidence that numerous pathogens have evolved mechanisms to specifically evade or subvert the IgA immune response is an indicator of the importance of IgA in protection against infection. These evasion mechanisms include proteins that bind specifically to IgA, inhibiting its action, as well as specific proteases that inactivate IgA through cleavage (10). Important examples of IgA-BPs include proteins Arp4 and Sir22 of group A Streptococcus, which are members of the M protein family (16, 17), and β-protein, an unrelated protein expressed by group B Streptococcus (18, 19). These streptococcal proteins interact with the IgA Cα2-Cα3 domain interface, which is essentially the same site recognized by FcαRI. Therefore, the IgA-BPs can inhibit IgA binding to FcαRI, as well as the triggering of the FcαRI-mediated respiratory burst (12). Another example of IgA-BP is represented by staphylococcal superantigen-like 7 (SSL7), a secreted toxin of Staphylococcus aureus which is able to bind monomeric IgA1 and IgA2 and also SIgA, competitively inhibiting FcαRI binding (20). Our finding that SIgA, the predominant antibody in the urinary tract secretions (10), is a crucial interactor with EsiB contributing to its immunomodulatory properties is of particular relevance to E. coli pathogenesis. Although we demonstrated that the cognate interaction of EsiB with SIgA is associated with a single peptide containing the E251 amino acid, we were not able to map the SIgA region bound by the antigen. Indeed, MS/MS analysis of a tryptic digest of purified SIgA immobilized on a chip and then incubated with EsiB revealed no interaction with isolated Fab or Fc portions, suggesting that a complete immunoglobulin is required for the binding (data not shown). Although the affinity constant of EsiB for SIgA is relatively low (K_d of approximately 10⁻⁷ M), it is conceivable that the avidity effect caused by the interaction of EsiB with multiple binding sites on SIgA may increase the local affinity, which was evaluated here as a 1:1 interaction. It can be difficult to predict whether during specific physiological events, affinity constants measured in vitro would correlate with the capacity of an antigen to sequester target molecules. However, we suggest that the interaction of EsiB with SIgA should at least imply high local concentrations of EsiB or the modulation of SIgA binding by other factors.

EsiB, unlike most bacterial IgA-BPs, cannot block cell surface binding of SIgA to FcαRI, which is in agreement with the antigen incapability of binding to the isolated Fc region of SIgA. However, we showed that EsiB was able to inhibit the antibody effector functions, such as superoxide generation and cellular migration, by interfering with the SIgA-triggered neutrophil activation. Indeed, EsiB interaction with SIgA was able to significantly decrease the tyrosine phosphorylation level of neutrophil PLC-γ, ERK1/2, and p38, which are all downstream components of the FcαRI signal transduction pathway that leads to cell activation. These data strongly suggest that EsiB can prevent productive FcαRI cross-linking by SIgA on the neutrophil plasma membrane, thus inhibiting the recruitment of immune cells at the infection site and the ROS-mediated killing of the pathogen.

To our knowledge, this study represents the first report of an E. coli SIgA-binding antigen that allows the bacterium to subvert host immune responses and avoid clearance by inhibiting neutrophil activation.

MATERIALS AND METHODS

Ethics statement.

The institutional review board of the Department of Health Service at Novartis Vaccines and Diagnostics (Siena, Italy) approved the study and the use of human samples from the volunteers. Written, informed consent was obtained from the adult participants.

Cells, antibodies, and reagents.

Polymorphonuclear neutrophils (PMNs) were purified from buffy coats from anonymous healthy donors (available from authorized blood banks) by density gradient centrifugation (400 × g for 30 min at room temperature) on Ficoll-Paque Plus (GE Healthcare), followed by centrifugation (250 × g for 10 min at 4°C) on a 3% (wt/vol) dextran solution (Sigma-Aldrich). After osmotic lysis of erythrocytes, cells were resuspended in RPMI 1640 (Invitrogen) supplemented with 10% fetal bovine serum (FBS) (Invitrogen) and incubated at 37°C in a humidified atmosphere with 5% CO₂.

Anti-EsiB polyclonal antibodies were raised in rabbits upon immunization with the purified recombinant protein. The anti-EsiB monoclonal antibody (MAb) was made from a hybridoma produced from a mouse immunized with highly purified EsiB. The antibody, which recognizes a linear epitope in the protein, was protein G purified from the hybridoma supernatant. Antiphosphotyrosine and antiactin MAbs were purchased from Millipore. Phosphospecific antibodies recognizing the phosphorylated active forms of ERK1/2 (T202/Y204) and p38 (T180/Y182) were from Cell Signaling Technology. Polyclonal anti-ERK2 and anti-p38 antibodies were obtained from Santa Cruz Biotechnology. Alkaline phosphatase-, peroxidase-, and FITC-labeled secondary antibodies were purchased from Dako. Alexa Fluor 488-labeled secondary antibody was from Invitrogen. FITC-labeled anti-human IgA antibody was obtained from Sigma-Aldrich.

Bacterial strains and culture conditions.

A total of 267 E. coli strains, including ExPEC (n = 119), InPEC (n = 117), and nonpathogenic (n = 31) strains, were analyzed (see Table S1 in the supplemental material). Genomic DNA was isolated using the GenElute bacterial genomic DNA kit (Sigma) according to the manufacturer’s instructions and kept at 4°C until further use. Uropathogenic E. coli (UPEC) strain CFT073 (serotype O6:K2:H1), isolated from the blood and urine of a patient with acute pyelonephritis (21), and isogenic mutant strain CFT073ΔesiB were used for real-time quantitative reverse transcription-PCR (RT-PCR) studies. Strains were cultured in Luria-Bertani (LB) broth (10 g/liter tryptone, 5 g/liter yeast extract, 10 g/liter NaCl) at 37°C with agitation and aeration, and approximately 10⁹ bacteria were collected after 2 h (log phase) and 5 h (stationary phase) for RNA isolation. E. coli DH5α-T1R (Invitrogen) was used for cloning purposes, and E. coli BL21(DE3) (Invitrogen) was used for expression of His-tagged fusion proteins. The clones carrying a specific cassette conferring antibiotic resistance were grown in the presence of kanamycin (50 µg/ml) or ampicillin (100 µg/ml).

Distribution analysis of the esiB gene.

To analyze the esiB gene prevalence, we performed a PCR screening approach, using primers EsiB_For, CGACTGGTTAGACAGGGATAAG, and EsiB_Rev, AGGAAAAGCAAAAGGCGAGG, designed in the external conserved region. For the distribution studies, the reaction mixture, containing a final volume of 25 ml and 100 µg of genomic DNA, was prepared using GoTaq Green Master Mix (Promega). The samples were subjected to 30 cycles of amplification in a thermal cycler (GeneAmp PCR system 9700; Applied Biosystems). Reaction conditions were 10 min of initial denaturation at 94°C, followed by 30 cycles of denaturation (30 s at 94°C), annealing (30 s for 55°C), and elongation (90 s at 72°C) and a final elongation at 72°C for 10 min. The expected PCR product was 1,770 bp. Additionally, BLAST analysis was performed on publicly available whole-genome sequences of E. coli strains (ftp://ftp.ncbi.nih.gov/genbank/genomes/Bacteria/; see Table S1 in the supplemental material) to detect the esiB gene.

Cloning, expression, and purification of EsiB.

The esiB gene, both with and without the predicted signal sequence, was amplified by PCR from the CFT073 genomic DNA template, cloned in the pET-21b vector (Novagen), and transformed in DH5α-T1R chemically competent cells for propagation. BL21(DE3) chemically competent cells were used for His-tagged protein expression. Purification of the recombinant protein, without signal peptide, was performed from the bacterial soluble fraction using nickel-affinity chromatography as already described (22, 23).

RNA purification and real-time quantitative RT-PCR.

Prior to RNA isolation, 2 volumes of RNA Protect bacterial reagent (Qiagen) was added to 1 volume of bacterial culture to immediately stabilize the in vivo gene expression profile in bacteria. For total RNA extraction, cells were disrupted by treatment with 15 mg/ml lysozyme (Sigma-Aldrich) and 2 mg/ml proteinase K (Qiagen) for 10 min at room temperature. RNA isolation was completed with the RNeasy minikit (Qiagen) according to the manufacturer’s instructions. RNA concentration and integrity were assessed by measurement of the A₂₆₀/A₂₈₀ ratios and electrophoretic analysis. Prior to RT-PCR, contaminating DNA was removed from RNA samples by treatment with RQ1 RNase-free DNase (Promega). RT-PCR was carried out on 500 ng total RNA using ImProm-II reverse transcriptase (Promega) and random primers (Promega) for first-strand cDNA synthesis.

Real-time quantitative PCR was performed using FastStart Universal SYBR Green Master (Rox) (Roche Diagnostics) in a LightCycler 480 II real-time PCR system (Roche Diagnostics). All samples were run in duplicate on 96-well optical PCR plates (Roche Diagnostics). The specific primers used to amplify cDNA fragments corresponding to EsiB mRNA and 16S rRNA were 5′ AAAGATGAGCAACAGGCCGCAATC 3′ (forward) and 5′ ATCACGTTCTACGCCCAAGCCATA 3′ (reverse) for EsiB and 5′ ACGTGCTACAATGGCGCATA 3′ (forward) and 5′ TCATGGAGTCGAGTTGCAGA 3′ (reverse) for 16S rRNA. After an initial denaturation at 95°C for 10 min, denaturation in the subsequent 40 cycles was performed at 95°C for 15 s, followed by primer annealing at 60°C for 30 s and a final extension at 72°C for 30 s. The specificity of the amplified fragments was demonstrated by the melting curve, where a single peak was observed for each sample amplified with EsiB and 16S rRNA primers. After amplification, data analysis was performed using LightCycler 480 1.5 software (Roche Diagnostics). For relative quantification of gene expression, the starting mRNA copy number of the unknown samples was determined using the comparative threshold cycle (ΔΔC_T) method, as previously described (24), and levels of the different transcripts were normalized to 16S rRNA, used as a housekeeping gene. For absolute quantification, serial dilutions of an external homologous standard (plasmid DNA carrying the cloned target sequence) were used to generate a standard curve, and data analysis was performed using the second derivative maximum method.

Measurement of serum IgG titers against EsiB.

Anti-EsiB immunoglobulin G (IgG) titers of patient sera were measured by enzyme-linked immunosorbent assay (ELISA). Briefly, Nunc-Immuno MaxiSorp 96-well plates (Thermo, Fisher Scientific) were coated overnight at 4°C with 0.5 µg of recombinant protein EsiB diluted in phosphate-buffered saline (PBS). After blocking with a 3% (wt/vol) polyvinylpyrrolidone (PVP) solution for 1 h at 37°C, plates were incubated with diluted patient sera for 2 h at 37°C. Bound IgG molecules were detected using alkaline phosphatase-conjugated goat anti-human IgG antibodies (Sigma-Aldrich), followed by the addition of a chromogen-substrate solution containing p-nitrophenylphosphate (Sigma-Aldrich). After 30 min, the reaction was stopped with 4 N NaOH and the absorbance at 405 nm was measured.

Binding of EsiB to SIgA (ELISA).

EsiB binding to human secretory immunoglobulin A (SIgA) was measured by ELISA. Briefly, Nunc-Immuno MaxiSorp 96-well plates (Thermo, Fisher Scientific) were coated overnight at 4°C with 1 µg of human colostrum IgA (Sigma-Aldrich) diluted in PBS. After blocking with a 3% (wt/vol) PVP solution for 2 h at 37°C, plates were incubated with serially diluted recombinant EsiB for 2 h at 37°C. Bound EsiB was detected using mouse anti-EsiB monoclonal antibodies and peroxidase-conjugated rabbit anti-mouse IgG antibodies (Dako), followed by the addition of a chromogen substrate solution containing o-phenylenediamine dihydrochloride (Sigma-Aldrich). After 20 min, the reaction was stopped with a 12.5% (vol/vol) H₂SO₄ solution and the difference between absorbance at 492 nm and absorbance at 630 nm was measured.

Binding of EsiB to SIgA (SPR).

The interaction between EsiB and secretory IgA antibodies was studied by surface plasmon resonance (SPR) using a BIAcore T200 instrument (GE Healthcare). Commercial secretory IgA antibodies (Sigma) were covalently immobilized by amine coupling on a carboxymethylated dextran sensor chip (CM-5; GE Healthcare). Amine-coupling reactions for immobilization of IgAs were performed using the purified commercial antibodies that have been previously exchanged in PBS buffer. SIgAs at ~10 µg/ml in 10 mM sodium acetate buffer, pH 5, were injected at 5 µl/min until ~1,800 response units (RU) was captured. The blank (reference) surface was constructed similarly, except that antibodies were omitted.

Running buffer containing phosphate-buffered saline (PBS) with 0.05% Tween 20 (pH 7.4) was used, and EsiB protein at a concentration range of 20 to 0.15 µM was injected. Injections were performed at 25°C with flow rates of 50 µl/min. Following each 60-s injection, a long dissociation time of 600 s allowed sensor chip surfaces to return to baseline without the need of a regeneration step. At least one concentration was included as replicate during each titration. SPR data were analyzed using the BIAcore T200 evaluation software (GE Healthcare). Titration experiments were performed in triplicate, and each sensorgram was fitted with the 1:1 Langmuir binding model, including a term to account for potential mass transfer, to obtain the individual k_on and k_off kinetic constants; the individual values were then combined to derive the single averaged K_d value reported.

Epitope mapping of SIgA-binding site on EsiB.

Epitope mapping experiments were carried out with recombinant EsiB and SIgA. To this aim, we adopted and extended an immunocapturing approach successfully used previously with monoclonal antibodies (25). The approach involves proteolytic digestion (using diverse proteases) of the target antigen prior to immunocapturing and subsequent analysis of antibody-bound peptides (containing the epitope) by mass spectrometry. In order to identify the region of interaction between SIgA and EsiB, we have used the Dynabeads antibody coupling kit (catalog no. 143.11D; Invitrogen). This kit allows covalent immobilization of antibodies (or other protein ligands such as functional enzymes, etc.) onto the surface of Dynabeads. Antibody-coupled Dynabeads can then be used for downstream assays and experiments like immunoassays, immune precipitation, coimmunoprecipitation of protein complexes, etc. Captured proteins or protein complexes are easily separated from the sample using the magnetic separation properties of Dynabeads. Magnetic separation facilitates washing, buffer changes, and elution. Secretory IgA was coupled onto the surface of Dynabeads, and these were then used to capture the epitope-containing peptide. Peptide mixtures of EsiB were obtained by partial digestion with trypsin and GluC, independently, in 100 mM ammonium bicarbonate buffer at 37°C for 3 h. The immunocapturing procedure was carried out as described previously (26). Briefly, a 25-µl suspension of SIgA-coated beads was washed twice with PBS using a magnet and resuspended to the initial volume. An 0.5-µl amount of protease inhibitor mixture (GE Healthcare) was added to the beads prior to the peptide mixture to avoid potential degradation of the antibodies. On addition of the peptide mixture, the sample was incubated for 30 min at room temperature with gentle tilting and rotation. After incubation, the beads were washed three times with 1 ml of PBS, and the bound peptide was then eluted with 50 µl of 0.2% trifluoroacetic acid (TFA). In order to obtain a suitable sample for mass spectrometry analysis, elute fractions were desalted and concentrated using C₁₈ ZipTips (Millipore). For the identification of the captured peptides, an ultrafleXtreme (Bruker Daltonics, Bremen, Germany) matrix-assisted laser desorption ionization–tandem time of flight (MALDI-TOF/TOF) mass spectrometer was used. For MALDI analysis, 1 µl of sample was mixed with the same volume of a saturated solution of α-cyano-4-hydroxy-trans-cinnamic acid matrix (0.5 mg/ml in water-acetonitrile-trifluoroacetic acid, 1%, 6:3:1) and spotted onto a MALDI target plate (Bruker). The drop was air dried at room temperature. MALDI mass spectra were recorded in the positive ion mode, and ion acceleration was set to 25 kV. All mass spectra were externally calibrated using a standard peptide mixture (Bruker); calibration was considered good when a value below 1 ppm was obtained. For peptide mass fingerprinting analysis, the Mascot search engine (Matrix Science) was used with the following parameters: two missed cleavage permission, 10-ppm measurement tolerance. Methionine oxidation was set as a variable modification. Common contaminants were removed using the contaminant database available in the Mascot search engine. Searches were carried out using the NCBInr NC_000908.2 database. For MS/MS analysis, measurement tolerance was set to 50 ppm for MS and to 0.2 Da for MS/MS. Positive identifications were accepted with a Mascot score higher than that corresponding to a P value of 0.05.

Binding of SIgA and EsiB to neutrophils.

To measure SIgA and EsiB binding to neutrophils, 5 × 10⁵ cells were incubated with different concentrations of SIgA or EsiB in Hanks’ balanced salt solution (HBSS) in 96-well culture plates for 20 min at 37°C. Cells were then washed and incubated for 30 min on ice with FITC-labeled anti-human IgA secondary antibody for SIgA detection or with anti-EsiB MAb followed by FITC-labeled secondary antibody for EsiB detection. For competition experiments, SIgA was preincubated with EsiB for 15 min at 37°C before addition to cells. Cell staining was analyzed by flow cytometry.

Confocal microscopy analysis of E. coli strains adhering to tissues.

Confocal imaging on tissues was performed on bladder sections derived from ExPEC-infected animals. In detail, the bladder was cut transversely, preserved in 10% formalin (pH 7.2), and embedded in paraffin. Tissue sections were cut and mounted onto slides. Samples were dewaxed and subjected to antigen retrieval before the immunofluorescence labeling. Staining of the sections was done by using a rabbit anti-EsiB serum followed by a donkey anti-rabbit IgG conjugated to rhodamine red X as a secondary antibody. Alexa Fluor dye-labeled phalloidins (Invitrogen) were used to stain the actin cytoskeleton according to the manufacturer’s instructions. Bacterial and cellular DNA was stained with 4′,6-diamidino-2-phenylindole (DAPI). The ProLong Gold reagent (Invitrogen) was used as liquid mounting. The CFT073 strain was detected using the polyclonal rabbit anti-CFT073 and the Alexa Fluor 488-conjugated goat anti-rabbit IgG secondary antibody. Tissues were labeled with wheat germ agglutinin (WGA) and conjugated with Alexa Fluor 647 (Invitrogen). The analysis was done with a Zeiss LSM 710 laser scanning microscope.

Respiratory burst measurement.

PMN respiratory burst activity was measured by chemiluminescence. Briefly, neutrophils were diluted to a final concentration of 10⁵ cells/ml in HBSS containing 0.1% human serum albumin (HSA). SIgA and EsiB-treated SIgA were then mixed with neutrophils. Following addition of luminol-balanced salt solution (LBSS), the oxidative burst was measured every 30 s for 60 min at 37°C by chemiluminescence using a luminometer (Berthold Technologies, Wildbad, Germany).

Chemotaxis assay.

To measure neutrophil chemotaxis (directed migration), bottom chambers of Transwell supports were filled with supernatants deriving from PMNs incubated for 20 min at 37°C with SIgA or EsiB-treated SIgA. Neutrophils (2.5 × 10⁵) were added to the upper chambers. After 1 h at 37°C, cells that had migrated toward the lower compartments were quantified by flow cytometry.

Cell activation, lysis, and immunoblot assays.

Neutrophil activation by FcαRI cross-linking was performed by incubating cells (5 × 10⁶) in HBSS with SIgA or EsiB-treated SIgA for 1 min at 37°C. Cells were recovered by centrifugation (16,000 × g for 30 s at 4°C), washed twice in PBS, and lysed in 1% (vol/vol) Triton X-100 in 20 mM Tris-HCl, pH 8.0, 150 mM NaCl, in the presence of a protease inhibitor cocktail (0.2 mg/ml sodium orthovanadate, 1 µg/ml pepstatin, 1 µg/ml leupeptin, 1 µg/ml aprotinin, and 10 mM phenylmethylsulfonyl fluoride). Equal amounts of proteins from each sample (measured using the Bio-Rad DC protein assay kit) were resolved by 10% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and transferred to 0.45-µm nitrocellulose filters (Bio-Rad). Prestained molecular mass markers (Invitrogen) were included in each gel. Immunoblotting was performed using primary antibodies and peroxidase-labeled secondary antibodies according to the manufacturer’s instructions and using a chemiluminescence detection kit (Thermo, Fisher Scientific). All blots were reprobed with loading control antibodies after stripping.

Statistical analysis.

Mean values, standard deviations, and the P values associated with two-tailed unpaired Student’s t test were calculated using the Microsoft Excel application. A level of P < 0.05 was considered statistically significant.

Protein structure accession number.

The crystal structure of EsiB was deposited in the PDB under accession no. 4BWR.

SUPPLEMENTAL MATERIAL

Figure S1

EsiB distribution by sequence type (ST). In red are represented the STs where c5321 is present. The figure includes all the sequence-typed strains of E. coli present in the database. Download

Figure S2

Immunoblot analysis, using serum of mice infected subcutaneously with the CFT073 strain (top) or preimmune serum (bottom), of EsiB recombinant protein. An unrelated protein was used as a negative control. Download

Figure S3

Kinetics of esiB expression in minimal medium. esiB transcript levels in the UPEC strain CFT073 were evaluated by real-time quantitative RT-PCR at different time points ranging from 1 to 7 h. The graph represents a typical experiment out of three performed with similar results. Download

Table S1

Multilocus sequence typing (MLST) and ancestral group analysis of the strains PCR positive for EsiB.