The Kingella kingae PilC1 MIDAS Motif Is Essential for Type IV Pilus Adhesive Activity and Twitching Motility

ABSTRACT

Kingella kingae is an emerging pathogen that has recently been identified as a leading cause of osteoarticular infections in young children. Colonization with K. kingae is common, with approximately 10% of young children carrying this organism in the oropharynx at any given time. Adherence to epithelial cells represents the first step in K. kingae colonization of the oropharynx, a prerequisite for invasive disease. Type IV pili and the pilus-associated PilC1 and PilC2 proteins have been shown to mediate K. kingae adherence to epithelial cells, but the molecular mechanism of this adhesion has remained unknown. Metal ion-dependent adhesion site (MIDAS) motifs are commonly found in integrins, where they function to promote an adhesive interaction with a ligand. In this study, we identified a potential MIDAS motif in K. kingae PilC1 which we hypothesized was directly involved in mediating type IV pilus adhesive interactions. We found that the K. kingae PilC1 MIDAS motif was required for bacterial adherence to epithelial cell monolayers and extracellular matrix proteins and for twitching motility. Our results demonstrate that K. kingae has co-opted a eukaryotic adhesive motif for promoting adherence to host structures and facilitating colonization.

INTRODUCTION

Kingella kingae is an emerging pediatric pathogen that is a leading cause of osteoarticular infections in the pediatric population (1–3). This Gram-negative bacterium colonizes the upper respiratory tract of young children, where it is usually present as a commensal organism. Carriage is a common occurrence, with approximately 10% of young children carrying the bacterium in their oropharynx at any given time (4–8). In some cases, K. kingae can breach the epithelial barrier, enter the bloodstream, and cause invasive diseases. Recent advances in culture-based and molecular diagnostics have identified K. kingae as a common cause of septic arthritis, osteomyelitis, and bacteremia in young children (1, 9–14). Adherence to epithelial cells represents the first step in K. kingae colonization of the oropharynx and is a prerequisite for invasive disease.

Previous studies have established that type IV pili and the pilus-associated PilC1 and PilC2 proteins are required for full-level K. kingae adherence to epithelial cells in vitro (15). Type IV pili are long, polymeric surface fibers produced by a wide range of Gram-negative and Gram-positive bacteria that can mediate numerous functions, including adherence, twitching motility, and natural competence (16). The fibers are assembled via polymerization of a major pilin subunit (PilA1 in K. kingae) at the inner membrane through the action of an assembly ATPase (PilF in K. kingae); the growing fiber extends through the periplasm and is secreted through an outer membrane secretin ring (PilQ in K. kingae), and it can be retracted though the secretin via the action of a retraction ATPase (PilT in K. kingae). Previous studies have demonstrated that K. kingae mutants with deletion of both pilC1 and pilC2 have a severe piliation defect and are non-adherent, incapable of twitching motility, and not naturally transformable (15, 17–19). Furthermore, the presence of type IV pilus fibers on the bacterial surface is necessary for the observed PilC-mediated phenotypes (15, 18, 19).

Additional studies using purified PilC1 and PilC2 have shown that these proteins directly interact with epithelial cells, with the adhesive region of the proteins located in their N-terminal domains (19). However, the specific binding motifs of PilC1 and PilC2 have not yet been determined. K. kingae PilC1 and PilC2 belong to a family of proteins which encompasses PilC1 and PilC2 in pathogenic Neisseria species, PilY1 in Pseudomonas aeruginosa, and PilY1 in Legionella pneumophila. This family is characterized by a structurally conserved β-propeller fold domain in the C-terminal region that is involved in pilus biogenesis (19, 20). K. kingae PilC1, L. pneumophila PilY1, and P. aeruginosa PilY1 all contain a predicted von Willebrand factor type A (vWFa) domain in the N-terminal region of the proteins (21). This domain is absent in K. kingae PilC2 and in the pathogenic Neisseria PilC1 and PilC2 proteins. The structural differences between K. kingae PilC1 and PilC2 are not surprising given their lack of similarity at the amino acid level, where they share only 8% identity and 16% similarity (15), contrasting with the pathogenic Neisseria PilC1 and PilC2 proteins, which are roughly 75% identical and 85% similar at the amino acid level (22–24).

The vWFa domain is frequently involved in multiprotein complexes and cell adhesion (25) and is so named for its presence in the von Willebrand factor glycoprotein, where it functions in platelet adhesion and blood clotting (26, 27). A common feature found in the vWFa domain in different proteins is the presence of a metal ion-dependent adhesion site (MIDAS) motif, which has been implicated in adherence in other molecules, most notably integrins (28–30). The MIDAS motif contains an aspartate residue followed by two serines, with any single amino acid in between each critical residue (DxSxS) and with a downstream threonine. The MIDAS motif is characterized as having a metal ion coordinated by oxygen atoms on the serines and water molecules bound by the aspartate (31). The metal ion is thought to be further stabilized by an acidic residue of the ligand, which enables the motif to mediate an adhesive interaction (28).

MIDAS motifs have been well studied in eukaryotic proteins (25, 32). Integrin MIDAS motifs have been shown to mediate integrin binding to various ligands, including intercellular cell adhesion molecules (ICAMs), vascular cell adhesion molecules (VCAMs), and extracellular matrix (ECM) proteins, including collagen I, collagen IV, laminin, and fibronectin (28, 32). Although MIDAS motifs are well understood in eukaryotes, they are not well characterized in prokaryotic organisms; however, these motifs have been identified in some bacterial proteins and have been suggested to play a role in bacterial colonization and interactions with host cells (33, 34).

In the present study, we identified and analyzed the predicted MIDAS motif located in the vWFa domain of K. kingae PilC1. Using site-directed mutagenesis, we generated a K. kingae mutant strain containing mutations in the critical residues of the putative adhesive motif. The K. kingae mutant strain exhibited defects in adherence to host epithelial cells and extracellular matrix proteins and in twitching motility. Our results demonstrate that K. kingae has adopted a eukaryotic adhesive motif for interacting with host molecules and persisting in the human host.

RESULTS

K. kingae PilC1 contains a predicted MIDAS motif that is not essential for surface piliation.

Given the presence of the vWFa domain in multiple PilC family members, we reasoned that this domain may play a key role in PilC protein function. Using the NCBI database, we identified a predicted MIDAS motif in the vWFa domain in the N-terminal region of the K. kingae PilC1 protein. The putative K. kingae PilC1 MIDAS motif contains the DxSxS critical residues and a downstream threonine at the N-terminal region of the protein. To investigate the possibility that the K. kingae PilC1 MIDAS motif is directly involved in mediating type IV pilus adhesive activity, we generated a mutant strain which contains point mutations at the critical residues of the PilC1 MIDAS motif (DxSxS). Using a primer-based site-directed mutagenesis approach, these residues were mutated to alanines (AxAxA) in strain KK03ΔpilC2-ErmPilC1 (Table 1), a pilC2 knockout strain which has an unmarked deletion of pilC2 and contains an erythromycin resistance cassette upstream of the pilC1 locus. The erythromycin cassette allowed selection for K. kingae transformants which incorporated the point mutations, yielding strain KK03ΔpilC2-ErmPilC1_AxAxA.

TABLE 1.

Strains and plasmids used in this study

Strain or plasmid	Description	Reference or source
Escherichia coli
XL-10 Gold	TetrΔ(mcrA)183 Δ(mcrCB-hsdSMR-mrr)173 endA1 supE44 thi-1 recA1 gyrA96 relA1 lac I [F′ proAB lacIqZΔM15 Tn10 (Tetr) Amy Camr]	Agilent
BL21(DE3)	E. coli B F− dcm ompT hsdS(rB− mB−) gal λ(DE3)	56
Kingella kingae
KK03	Naturally occurring spreading and corroding variant of septic arthritis clinical isolate 269–492	51
KK03 derivatives
ΔpilA1	KK03 with either aphA3- or tetM-marked pilA1 deletion	15, 19
ΔpilC1	KK03 with tetM-marked pilC1 deletion	18
ΔpilC2	KK03 with unmarked pilC2 deletion	19
ΔpilC1ΔpilC2	KK03 with tetM-marked pilC1 deletion and an unmarked pilC2 deletion	19
ΔpilT	KK03 with ermC-marked pilT deletion	19
ΔpilC1ΔpilC2ΔpilT	KK03 with tetM-marked pilC1 deletion, unmarked pilC2 deletion, and ermC-marked pilT deletion	19
ΔpilC2-ErmPilC1	KK03 pilC2 deletion strain with ermC insertion upstream of WT pilC1	18
ΔpilC2-ErmPilC1AxAxA	ΔpilC2-ErmPilC1 with PilC1 D90A S92A S94A mutations	This work
Plasmids
pET22b	Protein expression vector	MilliporeSigma
pET22b/PilC1	For expression of 6×HisPilC1	19
pET22b/PilC1AxAxA	For expression of 6×HisPilC1AxAxA	This work
pUC19/ErmpilC1	For generating MIDAS point mutations	18
pUC19/ErmpilC1AxAxA	For introducing PilC1D90A S92A S94A	This work

With the MIDAS mutant strain in hand, we first tested the effect of the MIDAS motif mutation on K. kingae PilC1 production. As shown in Fig. 1A, the level of PilC1 in the MIDAS mutant strain (KK03ΔpilC2-ErmPilC1_AxAxA) was comparable to the level in the isogenic control strain (KK03ΔpilC2-ErmPilC1). Lower levels of PilC1 were observed in the MIDAS mutant strain and the control strain compared with the wild-type strain, potentially due to placement of the erythromycin marker upstream of the pilC1 promoter. However, total levels of surface piliation based on the quantity of the major pilin subunit PilA1 in sheared pilus preparations were comparable in the MIDAS mutant, control, and wild-type strains (Fig. 1A, middle gel image labeled “PilA1”), indicating that the point mutations do not affect PilC1-mediated pilus expression. Western blot analysis with antiserum GP-25 directed against the PilC2 N-terminal region confirmed that PilC2 was absent from the strains used to evaluate only PilC1-mediated phenotypes. Western blot analysis of whole-cell lysates with an antiserum specific for PilA1 demonstrated that all strains except for KK03 ΔpilA1 produced similar levels of PilA1. Western analysis with an antiserum against K. kingae GAPDH (glyceraldehyde-3-phosphate dehydrogenase) was used as a loading control to demonstrate that similar amounts of bacteria were subjected to the pilus preparation procedure. As shown in Fig. 1B, pilus fibers in the MIDAS mutant strain were morphologically similar to those in the wild-type strain KK03 when viewed by transmission electron microscopy (Fig. 1B). Taken together, these data indicate that mutation of the MIDAS motif does not affect surface piliation.

FIG 1.

The Kingella kingae PilC1 metal ion-dependent adhesion site (MIDAS) motif is not required for surface piliation. (A) Pilus preparations and whole-cell lysates of K. kingae strains KK03, KK03ΔpilA1, KK03ΔpilC1, KK03ΔpilC2, KK03ΔpilC1ΔpilC2, KK03ΔpilC1ΔpilC2ΔpilT, KK03ΔpilC2-ErmPilC1, and KK03ΔpilC2-ErmPilC1_AxAxA were boiled and separated using SDS-PAGE. The PilA1 pilin monomer band in the sheared pilus preparation was stained with Coomassie blue (samples labeled “PilA1”). Western blotting was used to detect PilC1 and PilC2 in the sheared pilus preparations and PilA1 and GAPDH (glyceraldehyde-3-phosphate dehydrogenase) in the whole-cell lysates. (B) Morphology of type IV pili in strains KK03 and KK03ΔpilC2-ErmPilC1_AxAxA was examined using transmission electron microscopy following negative staining with uranyl acetate. Scale bar = 500 nm.

The MIDAS motif is required for K. kingae PilC1 binding to epithelial cells and ECM proteins.

To assess the role of the PilC1 MIDAS motif in K. kingae adherence to epithelial cells, we examined bacterial adherence to monolayers of Chang epithelial cells of HeLa origin. As shown in Fig. 2, strain KK03ΔpilC2-ErmPilC1_AxAxA was non-adherent in these assays. To further investigate the role of the MIDAS motif in PilC1 binding activity, we purified recombinant PilC1 and PilC1_AxAxA and then examined protein binding to epithelial cell monolayers. PilC1_AxAxA showed a significant defect in binding to epithelial cells (Fig. 3A and B) compared to PilC1, demonstrating that the MIDAS motif is essential for PilC1 binding to epithelial cell monolayers. As shown in Fig. 3C and Fig. 3D, PilC1 and PilC1_AxAxA displayed similar circular dichroism curves, indicating that the point mutations did not grossly affect the global protein structure. PilC1 binding affinity decreased significantly in the presence of the metal ion chelator EDTA, as evidenced by the increased calculated K_d (dissociation constant; Fig. 3B). The binding defect observed in the presence of EDTA was reversed by the addition of MgCl₂, indicating that the presence of a divalent cation is required for PilC1 binding (Fig. 3A and B). Taken together, these data demonstrate that the PilC1 MIDAS motif and metal ions are essential for purified PilC1 binding to epithelial cells.

FIG 2.

The K. kingae PilC1 MIDAS motif is essential for bacterial adherence to epithelial cells. K. kingae strains KK03, KK03ΔpilA1, KK03ΔpilC2, KK03ΔpilC1ΔpilC2, KK03ΔpilC1ΔpilC2ΔpilT, KK03ΔpilC2-ErmPilC1, and KK03ΔpilC2-ErmPilC1_AxAxA were added to epithelial cell monolayers and assessed for adherence. Percent adherence was calculated based on the ratio of recovered bacteria to the inoculum. Error bars represent standard error of the mean, n = 3. *, P < 0.05 as determined by a one-way analysis of variance (ANOVA) using Bonferroni’s correction for multiple comparisons.

FIG 3.

The K. kingae PilC1 MIDAS motif and metal ions are essential for purified PilC1 binding to epithelial cells. Purified PilC1 and PilC1_AxAxA were diluted in 50 mM Tris-HCl, 50 mM Tris-HCl with 10 mM EDTA (PilC1 + EDTA), or 50 mM Tris-HCl with 10 mM EDTA and 12 mM MgCl₂ (PilC1 + EDTA + MgCl₂) at increasing concentrations ranging from 2.06 to 205.55 nM. (A) Proteins were added to epithelial cell monolayers, and binding was detected by ELISA using polyclonal antiserum CHP-GP7 against PilC1. (B) Nonlinear regressions were fitted using the GraphPad Prism one site-specific binding model fit to total data from 3 independent runs, and the apparent dissociation constant (K_d) was generated based on the regression. Error bars represent standard error of the mean, n = 3. ns, not significant. *, P < 0.05 as determined by a one-way ANOVA of K_d comparisons to PilC1 using Bonferroni’s correction for multiple comparisons. Circular dichroism curves were generated for purified recombinant PilC1 (C) or PilC1_AxAxA (D). Curves displayed are representative images of CD data collected from three separate batches of protein purified on different days.

In previous work, we demonstrated that PilC1 promotes K. kingae adherence to ECM proteins (19). To investigate the role of the PilC1 MIDAS motif in adherence to ECM proteins, we examined bacterial adherence to plates coated with collagen I, collagen IV, laminin, or fibronectin. Similar to our observations with epithelial cells, PilC1-mediated adherence to these ECM proteins was abolished in the MIDAS mutant strain, indicating that the PilC1 MIDAS motif is required for K. kingae adherence to ECM proteins (Fig. 4A to D).

FIG 4.

The K. kingae PilC1 MIDAS motif is essential for bacterial adherence to extracellular matrix proteins. K. kingae strains KK03, KK03ΔpilA1, KK03ΔpilC2, KK03ΔpilC1ΔpilC2, KK03ΔpilC1ΔpilC2ΔpilT, KK03ΔpilC2-ErmPilC1, and KK03ΔpilC2-ErmPilC1_AxAxA were added to tissue culture-treated plates coated with collagen I (A), collagen IV (B), laminin (C), or fibronectin (D) and assessed for adherence. Percent adherence was calculated based on the ratio of recovered bacteria to the inoculum. Error bars represent standard error of the mean, n = 3. *, P < 0.05 as determined by a one-way ANOVA using Bonferroni’s correction for multiple comparisons.

The K. kingae MIDAS motif is required for twitching motility.

Twitching motility is a type IV pilus-mediated process that involves sequential pilus extension, adhesion to substrate, and pilus retraction (35–40) and can be quantified by measuring bacterial spread using agar stab assays on chocolate agar plates. As shown in Fig. 5A and B, the MIDAS mutant strain was completely deficient in twitching motility, indicating that the K. kingae PilC1 MIDAS motif is also required for twitching motility.

FIG 5.

The K. kingae PilC1 MIDAS motif is essential for twitching motility. (A) K. kingae strains KK03, KK03ΔpilA1, KK03ΔpilC2, KK03ΔpilC1ΔpilC2, KK03ΔpilC1ΔpilC2ΔpilT, KK03ΔpilC2-ErmPilC1, and KK03ΔpilC2-ErmPilC1_AxAxA were added to chocolate agar motility plates, and the zone of twitching motility was stained with crystal violet. (B) Twitching motility was quantified by measuring the diameter of bacterial spread. Error bars represent standard error of the mean, n = 3. *, P < 0.05 as determined by a one-way ANOVA using Bonferroni’s correction for multiple comparisons.

DISCUSSION

Previous work has established a role for K. kingae type IV pili and the pilus-associated proteins PilC1 and PilC2 in bacterial adherence to epithelial cells and twitching motility (15, 18, 19). However, the molecular mechanisms of PilC1 and PilC2 involvement in these processes have remained unknown. In this study, we found that the K. kingae PilC1 MIDAS motif is essential for PilC1-mediated bacterial adherence to epithelial cell monolayers and ECM proteins. In addition, we observed that the MIDAS motif and metal ions are required for purified PilC1 binding to epithelial cells. We also observed that mutation of the MIDAS motif eliminated twitching motility. Our data from this work demonstrate that the K. kingae PilC1 MIDAS motif is essential for PilC1 adhesive activity.

MIDAS motifs have been well-studied in integrin molecules and have been shown to work cooperatively with metal ions to regulate integrin function and mediate binding to ECM proteins such as collagen I, collagen IV, fibronectin, and laminin (41). Integrin proteins are heterodimers, containing an α and a β chain, which function to promote cell-cell and cell-ECM interactions. Some α chains contain an extra stretch of ~200 amino acids called the inserted or “I” domain (32). These integrins have been shown to directly bind to motifs in collagens using the perfectly conserved MIDAS motif of the αI domain (42). In addition to the MIDAS motifs of integrin αI domains, all integrin β chains contain either a perfect or imperfect MIDAS motif, with perfect motifs retaining the critical residues DxSxS and imperfect motifs having at least one of those residues substituted with a residue of the same charge (25). In αI-less integrins, the ligand is thought to bind at the interface of the α and β chains, with the MIDAS motif playing a key role in facilitating this binding (32). Integrin MIDAS motifs are thought to coordinate a divalent cation together with an acidic residue of the ligand to mediate the adhesive interaction (28). X-ray crystallography studies have shown that divalent cations such as Mn²⁺, Mg²⁺, and Zn²⁺ can facilitate adhesion through coordination of the metal ion, while bulkier ions such as Ca²⁺ do not fit as well in the binding pocket and are not preferred by the motif (42, 43). Similar to integrin molecules, purified recombinant K. kingae PilC1 protein was able to utilize the MIDAS motif and Mg²⁺ to bind. Further analyses are necessary to determine whether other divalent cations in addition to Mg²⁺ can promote PilC1 MIDAS-mediated adhesive interactions. Given the similarities between PilC1-mediated binding and integrin-binding, we speculate that the K. kingae PilC1 MIDAS motif may have evolved to imitate an integrin molecule to gain access to these host cell receptors while also retaining the ability to promote twitching motility. Although the K. kingae PilC2 protein also promotes twitching motility, it does not contain a predicted MIDAS motif or mediate adherence to ECM proteins (18, 19), suggesting that both PilC1 and PilC2 evolved to retain the ability to promote twitching motility but to mediate adherence to difference substrates.

Laminin and collagen IV are present in the epithelial cell basement membrane in the oropharynx, where K. kingae resides as a commensal organism (4, 44–46). The K. kingae transition from being a commensal in the oropharynx to an invasive bacterium is not well understood. Previous studies have shown that invasive disease frequently develops concurrently with or after hand-foot and mouth disease (47) or infection with varicella virus (48), herpes simplex virus (48, 49), or rhinovirus (50). In addition, K. kingae secretes a broadly active RTX toxin, which has been shown to be cytotoxic to epithelial cells in vitro (51). Beneath the epithelial cell basement membrane lies the interstitial connective tissue layer, comprised of collagen I (45, 46). In the event of damage to the epithelial cells, which may happen as a consequence of viral coinfection or RTX-mediated cytotoxicity, the basement membrane and connective tissue may become exposed, along with potential receptors such as ECM proteins. With this information in mind, we speculate that K. kingae utilizes the PilC1 MIDAS motif to facilitate invasion when epithelial cells become damaged. Collagen I is also a major component of bones and joints, sites of the most common K. kingae invasive diseases, osteomyelitis and septic arthritis. Given the established role of the MIDAS motif in adherence to ECM proteins and the location of ECM proteins below epithelial cells and within bone and joints, it is possible that the PilC1 MIDAS motif plays a role in facilitating bacterial invasion into the bloodstream and subsequent seeding of sites of invasive disease.

MIDAS motifs are not well-characterized in prokaryotes; however, the limited studies characterizing the MIDAS motif in bacterial species suggest a role for this motif in bacterial adherence and colonization (33, 34). Using a catheter-associated urinary tract infection model, Nielsen et al. (33) showed that Enterococcus faecalis utilizes a MIDAS motif in the EbpA pilus subunit to colonize host tissue. Interestingly, Flores-Mireles et al. (52) showed that immunizing mice with the MIDAS-containing portion of the EbpA protein, but not other Ebp pilus components or the EbpA protein lacking the MIDAS motif domain, prior to bacterial challenge was able to protect against E. faecalis catheter-associated urinary tract infections. Although MIDAS motifs are present in many proteins, the proteins which utilize them for interaction with a ligand exhibit considerable specificity, which is conferred by the residues surrounding the MIDAS motif, rather than the MIDAS motif residues coordinating the metal ion (53). Therefore, incorporating protein domains containing MIDAS motifs may prove an effective vaccine strategy for preventing bacterial binding to host cells, and the role of the K. kingae PilC1 MIDAS motif in PilC1-mediated binding makes this region an attractive target for potential future vaccine development. However, further structural characterization of bacterial MIDAS motif-containing domains and immunogenicity studies in animals will be necessary to determine whether they are immunogenic and to exclude potential antibody cross-reactivity with eukaryotic proteins containing MIDAS motifs.

In this study, we show that the K. kingae PilC1 MIDAS motif is also essential for twitching motility. We propose two hypotheses for the role of the MIDAS motif in this dynamic process. Because the process of twitching motility requires sequential rounds of pilus extension, adherence to a substate, and pilus retraction, it is possible that pilus fibers containing PilC1 with a mutated MIDAS motif are defective in their ability to attach to a surface after pilus extension. In this scenario, the extended fibers are not adhered to a surface, resulting in an inability of the organism to generate tension on the fibers, precluding surface motility. An alternative possibility is that the MIDAS motif mutation results in a defect in pilus retraction, leading to ablated twitching motility. A hallmark of type IV pilus retraction proficiency in K. kingae is the ability to internalize exogenous DNA in a process known as natural transformation. We have previously shown that a mutant lacking PilC2 and missing the N-terminal domain of PilC1 (and thus deficient in the MIDAS motif) retains the ability to take in exogenous DNA, resulting in transformation, albeit at a reduced level compared to a strain with full-length PilC1 (19). In contrast, transformation is completely ablated in a pilT retraction ATPase mutant (19). Thus, the MIDAS motif-containing N-terminal domain of PilC1 is not absolutely required for pilus retraction.

These data establish a key role for the K. kingae PilC1 MIDAS motif in type IV pilus adhesive activity and demonstrate that K. kingae has adopted a eukaryotic adhesive motif to promote its adherence to host cells and twitching motility, processes which facilitate colonization and persistence in the host.

MATERIALS AND METHODS

Generation of K. kingae strains and the PilC1 MIDAS mutant.

K. kingae gene disruptions and mutations were generated as described previously (15, 18, 54). Plasmid-based disruption and directed mutagenesis constructs were generated in Escherichia coli. Point mutations were introduced using a QuikChange Lightning Mutagenesis kit (Agilent, Santa Clara, CA) for site-directed mutagenesis. Mutations were introduced into K. kingae using natural transformation of linearized plasmid DNA followed by selection for mutants on chocolate agar plates with the appropriate antibiotic. Mutations were confirmed by genomic DNA preparation of putative mutant strains followed by PCR amplification of the mutation and evaluation by Sanger sequencing.

To generate K. kingae strain KK03ΔpilC2-ErmPilC1_AxAxA, plasmid pUC19/ErmpilC1 was subjected to site-directed mutagenesis with the primers MIDAS sense (5′-CCCAATATTATGCTACTATTGGCGGACGCAGGAGCTATGGGTGCTCAAGTTCCAGGC-3′) and MIDAS antisense (5′-GCCTGGAACTTGAGCACCCATAGCTCCTGCGTCCGCCAATAGTAGCATAATATTGGG-3′) to introduce point mutations resulting in PilC1 containing D90A, S92A, and S94A mutations. The mutagenesis reaction was transformed into E. coli XL-10 Gold chemically competent cells followed by the selection of transformants using LB agar plates with 100 μg/mL ampicillin. The plasmid was purified, and the constructs were confirmed using Sanger sequencing. Plasmid pUC19/ErmpilC1_AxAxA was introduced into K. kingae strain KK03, and transformants were recovered by selection with erythromycin using the selection marker upstream of mutant pilC1. An internal fragment of pilC1_AxAxA was amplified and Sanger-sequenced, demonstrating that only MIDAS motif mutant-specific sequence was present in this strain.

Pilus preparations.

Pilus preparations were performed as described previously (19). Bacteria were grown at 37°C, 5% CO₂ for 20 h on chocolate agar plates, swabbed from the plate, and resuspended in 12 mL phosphate-buffered saline (PBS) to an optical density at 600 nm (OD₆₀₀) of 0.8. Samples were vortexed at full speed for 1 min and centrifuged at 4,000 × g for 30 min to pellet bacteria. A total of 10 mL of the bacteria-free supernatant was subjected to 20% ammonium sulfate precipitation on ice for 2 h. Precipitated pili were collected via centrifugation at 20,000 × g for 20 min and resuspended in 1× SDS-PAGE loading buffer. To visualize the level of piliation, pilus preparations were separated on 15% SDS-PAGE gels and stained with Coomassie blue to examine the major pilin subunit (PilA1) band. To detect PilC1 and PilC2, pilus preparations were separated using 7.5% SDS-PAGE and transferred to nitrocellulose. PilC1 was detected with a rabbit antiserum generated against its N-terminal region (Rab 128), and PilC2 was detected with a guinea pig antiserum generated against its N-terminal regions (GP-25). As pilus preparation controls, whole-cell lysates of the bacterial pellets following shearing of the pili were prepared by sonication. These samples were separated on 15% SDS-PAGE gels and transferred to nitrocellulose. PilA1 was detected with guinea pig antiserum GP-65 (19), and GAPDH was detected with GP-22 (19) as a loading control. Gel and blot images were acquired using a Syngene G:Box gel documentation system (Frederick, MD).

Transmission electron microscopy.

Negative-staining transmission electron microscopy was carried out as described previously (51, 55). Bacteria were resuspended in 200 mM ammonium acetate and absorbed onto Formvar carbon-coated copper mesh grids for 1 min. The grids were washed with 20 μL distilled water twice before being stained with 2% uranyl acetate. The grids were dried before being imaged using an FEI Tecnai T12 electron microscope.

Eukaryotic cell lines.

Chang epithelial cells (Wong-Kilbourne derivative [D] of Chang conjunctiva, HeLa origin; ATCC CCL-20.2) were grown at 37°C, 5% CO₂ in tissue culture treated plates in medium as described previously (51).

Quantitative bacterial adherence assays.

Quantitative adherence assays were performed as described previously (15, 18, 54). For determining adherence to Chang epithelial cells, epithelial cells were seeded into 24-well tissue culture-treated plates and incubated overnight at 37°C, 5% CO₂. The following morning, cells were fixed with 2% glutaraldehyde in 0.2 M sodium phosphate buffer (pH 7.4) and washed with Tris-buffered saline (TBS). For determining adherence to extracellular matrix proteins, Cellware plates coated with either collagen I, collagen IV, fibronectin, or laminin were purchased from Corning. ECM plates were left at room temperature to warm for 1 h prior to inoculation with bacteria. Bacteria were grown at 37°C overnight on chocolate agar plates, swabbed from the plate, and resuspended in brain-heart infusion (BHI) medium to an OD₆₀₀ of 0.8. The resuspended bacteria were added to 24-well tissue culture-treated plates either seeded with epithelial cell monolayers or coated with ECM proteins containing 300 μL of tissue culture medium. The plates were incubated at 37°C, 5% CO₂ for 25 min before being washed with PBS to remove unbound bacteria. A volume of 100 μL 0.05% trypsin was added to the plates, followed by incubation at 37°C, 5% CO₂ for 20 min to facilitate bacterial recovery. Recovered bacteria were plated onto chocolate agar, and percent adherence was calculated based on the ratio of recovered bacteria to the inoculum.

Generation of recombinant PilC1_AxAxA expression construct.

Plasmid pET22/PilC1 (19) was subjected to site-directed mutagenesis using the QuikChange Lightning Mutagenesis kit and primers MIDAS sense and MIDAS antisense to introduce the directed mutations, generating plasmid pET22/PilC1_AxAxA. The reaction was transformed into E. coli XL-10 Gold chemically competent cells followed by the selection of transformants using LB agar plates with 100 μg/mL ampicillin. The plasmid was purified, and the constructs were confirmed using Sanger sequencing. The plasmid was introduced into E. coli BL21 by electroporation, and transformants were recovered by selection with LB agar plates with 100 μg/mL ampicillin.

PilC1 and PilC1_AxAxA protein purification.

Overnight cultures of E. coli BL21 containing pET22/PilC1 or pET22/PilC1_AxAxA were back-diluted 1:200 in LB broth with 100 μg/mL ampicillin, and protein expression was induced upon reaching an OD₆₀₀ of 0.4 for 3 h at 30°C using 0.04 mM isopropyl β-d-1-thiogalactopyranoside (IPTG). The induction cultures were centrifuged at 6,700 × g for 20 min, and the supernatant was removed. The cell pellet was resuspended in a buffer containing 50 mM Tris-HCl, 5 mM EDTA, and 10 mM NaCl at pH 8.0 at a ratio of 3 mL buffer to 1 g of cells. AEBSF [4-(2-aminoethyl)benzenesulfonyl fluoride hydrochloride] was added to a final concentration of 1 mM. For each 1 g of cells, 0.8 mg of lysozyme was added. The solution was mixed and placed in a 37°C water bath until viscous. The solution was sonicated 3× at 25% amplitude for 30 s using a QSonica Q500 sonicator to shear DNA. The solution was centrifuged at 39,191 × g for 30 min, and the supernatant was discarded. Cells were resuspended in a buffer containing 20 mM Na₂HPO₄, 20 mM NaCl, 5 mM EDTA, and 25% sucrose at pH 7.2 at a ratio of 3 mL buffer to 1 g of cells. The same amount of AEBSF was added as previously. For each 1 mL of solution, 10 μL of Triton X-100 was added. The solution was mixed and centrifuged at 48,384 × g for 20 min, and the supernatant was discarded. The pellets were solubilized using 50 mM Tris-HCl, 40 mM imidazole, 8 M urea, and 1 mM β-mercaptoethanol at pH 8.0 for several hours at 37°C, adding more buffer until no more pellet would dissolve. The solubilized pellet was centrifuged at 48,384 × g for 20 min. The remaining supernatant was filtered using vacuum filtration, and 6× histidine-tagged PilC1 or PilC1_AxAxA was purified from the supernatant using affinity chromatography over an Ni-NTA column. The column was equilibrated to the previously mentioned solubilization buffer, and supernatants were added to the column and incubated at room temperature on a rotator for 3 h to facilitate protein binding. The column was washed with 20 mL of the previously mentioned solubilization buffer. Proteins were eluted with 50 mM Tris-HCl, 500 mM imidazole, 8 M urea, and 1 mM β-mercaptoethanol at pH 8.0. The samples were concentrated over a 100,000-Da molecular weight cutoff filter. The samples were then dialyzed in 50 mM Tris-HCl (pH 8.5) at 4°C for 24 h. The buffer was removed and replaced with fresh buffer and samples were dialyzed at 4°C for an additional 24 h. Protein samples were stored at 4°C in 50 mM Tris-HCl (pH 8.5).

Cell-based protein-binding assays.

A total of 3.24 × 10⁴ Chang epithelial cells were seeded into 96-well tissue culture-treated plates and incubated overnight at 37°C, 5% CO₂. The cells were then fixed with 2% glutaraldehyde in sodium phosphate buffer and washed three times with TBS. Monolayers were blocked using 2% dry milk powder in PBS for 1.5 h at 37°C. Protein dilutions ranging from 2.06 to 205.55 nM were prepared in 50 mM Tris-HCl (pH 8.5). The blocking buffer was removed, and the diluted proteins were added to either tissue culture treated plates only or tissue culture treated plates coated with epithelial cell monolayers. Protein-loaded plates were incubated at 37°C for 3 h before being washed four times with PBS. Polyclonal antiserum to PilC1 in 2% dry milk powder/PBS was added to the plates at 1:500. Plates were incubated at 37°C for 45 min before being washed four times with PBS. A secondary antibody conjugated to horseradish peroxidase in 2% dry milk powder/PBS was added to the plates at 1:2,000. Plates were incubated at 37°C for 45 min before being washed four times with PBS. 3,3′,5, 5′-Tetramethylbenzidine ELISA peroxidase substrate reagent was added to the plates for 11 min, and the absorbance was measured at 655 nm. Background signal to the plate was subtracted from total adherence detected to epithelial cell monolayers to generate specific binding data.

Twitching motility assays.

Twitching motility assays were performed as described previously (18). Strains were suspended to OD₆₀₀ = 0.8 in BHI medium. A 1-μL volume of the bacterial suspension was stab-inoculated into the center of a chocolate agar motility plate (chocolate agar with 1% agar) to the plate-agar interface using a pipette tip. Plates were incubated at 37°C, 5% CO₂ for 3 days, the chocolate agar was then carefully removed, and the zone of bacterial spread was stained with crystal violet. The diameter of the crystal violet-stained bacterial spread was measured in millimeters.

Circular dichroism.

PilC1 and PilC1_AxAxA proteins were diluted to a final concentration of 50 μg/mL in 16 mM Tris-HCl (pH 8.5) and analyzed using a Jasco J-810 spectropolarimeter, using a 0.1-cm cuvette at a wavelength range of 250 to 190 nm. The spectropolarimeter was blanked using 16 mM Tris-HCl (pH 8.5) prior to taking measurements. The parameters were set to 1-nm data pitch, standard sensitivity, 1-s data integration time (DIT), 1-nm bandwidth, immediate start mode, 20-nm/min scan speed, and 6 total accumulations.