Multi-functional analysis of Klebsiella pneumoniae fimbrial types in adherence and biofilm formation

Abstract

Klebsiella pneumoniae is an opportunistic pathogen frequently associated with nosocomially acquired infections. Host cell adherence and biofilm formation of K. pneumoniae isolates is mediated by type 1 (T1P) and type 3 (MR/K) pili whose major fimbrial subunits are encoded by the fimA and mrkA genes, respectively. The E. coli common pilus (ECP) is an adhesive structure produced by all E. coli pathogroups and a homolog of the ecpABCDE operon is present in the K. pneumoniae genome. In this study, we aimed to determine the prevalence of these three fimbrial genes among a collection of 69 clinical and environmental K. pneumoniae strains and to establish a correlation with fimbrial production during cell adherence and biofilm formation. The PCR-based survey demonstrated that 96% of the K. pneumoniae strains contained ecpA and 94% of these strains produced ECP during adhesion to cultured epithelial cells. Eighty percent of the strains forming biofilms on glass produced ECP, suggesting that ECP is required, at least in vitro, for expression of these phenotypes. The fim operon was found in 100% of the strains and T1P was detected in 96% of these strains. While all the strains examined contained mrkA, only 57% of them produced MR/K fimbriae, alone or together with ECP. In summary, this study highlights the ability of K. pneumoniae strains to produce ECP, which may represent a new important adhesive structure of this organism. Further, it defines the multi-fimbrial nature of the interaction of this nosocomial pathogen with host epithelial cells and inert surfaces.

Introduction

Klebsiella pneumoniae is a Gram-negative bacterium belonging to the Enterobacteriaceae family, often found in a variety of environmental niches.¹K. pneumoniae is an important nosocomial pathogen involved in diverse diseases such as pneumonia, urinary tract infections, bacteremia, and wound infections.¹K. pneumoniae infections are of great concern undoubtedly due to their emergence after the widespread use of antibiotics, particularly in nosocomial settings. In the last three decades, several nosocomial outbreaks of infections caused by multidrug-resistant K. pneumoniae have been reported.^2-5K. pneumoniae produces several structures that are essential for virulence, including pili that aid in the initial colonization of the host and capsular polysaccharides that protect the organism from phagocytosis, complement and inhibit macrophage differentiation.^1,6,7 Adherence of K. pneumoniae to eukaryotic epithelial cells is mainly attributed to two major adhesive pili structures, the mannose-sensitive type 1 pili (T1P) composed of a major fimbrial FimA subunit and a minor tip adhesin FimH, and the mannose-resistant type 3 pili (T3P or MR/K), composed of the major pilus subunit MrkA and the minor tip adhesin MrkD.^8-10 Other less studied K. pneumoniae adhesins are a third fimbrial structure named KPF-28, which was described in 83% of multi-drug resistant French isolates¹¹ and the non-fimbrial CF-29K antigen.^2,12

T1P are wide spread among members of the Enterobacteriaceae and their role in the pathogenesis of human urinary tract infections (UTI) caused by Escherichia coli and in experimental murine UTIs by K. pneumoniae is well established.^13-15 T1P mediate binding to mannose-containing receptors on epithelial cells of the urogenital tract and trachea, yeast cells and guinea pig erythrocytes.^16-18 In contrast, the MR/K pili adhesive functions are independent of D-mannose and are thought to be produced by most K. pneumoniae strains and to promote adherence to tracheal epithelial cells, renal tubular cells, extracellular matrix proteins, basement membranes of lung tissue and to aid in biofilm formation.^19-21 The mrkD gene is prevalent among K. oxytoca strains, but is rare in K. pneumoniae strains.^21,22 The MrkD tip adhesin is needed for tannic acid-dependent hemagglutination, to mediate binding in vitro to eukaryotic tissues, and for adherence to extracellular matrices.²³ However, mrkD⁻ strains can still bind to host cells and form biofilms, suggesting that MrkA is enough to facilitate bacterial interactions leading to the efficient development of biofilms and/or that other yet unrecognized adhesins exist.²⁰

Meningitis-associated E. coli (MENEC) strains produce a fimbrial structure called Mat (meningitis-associated temperature-dependent fimbriae) at non-physiological temperatures (e.g., 26°C) and were recently shown to be required for biofilm formation.^24,25 We demonstrated in a series of reports that these fimbriae are also produced at 37°C by most E. coli, including human and animal pathogenic strains, as well as fecal commensal strains.²⁶ Given the widespread nature of the genes encoding these pili among E. coli pathogroups we proposed the name E. coli common pilus or ECP.²⁶ The ECP is encoded by the chromosomal ecpRABCDE operon, and is composed of a major pilin subunit of 21 kDa protein called EcpA, which bears no biochemical resemblance to any other known pilus protein. The ecpR gene codes for a transcriptional activator, EcpR, which autoregulates itself and activates transcription of the entire operon including, the major pilin ecpA gene.^26-28 This pilus has been reported to aid in cell adherence of enterohemorrhagic (EHEC), enteropathogenic (EPEC) and enteroaggregative (EAEC) E. coli and is produced by a high percentage of enterotoxigenic (ETEC) strains.^26,28-30 A recent study showed that ECP plays a key role in bacteria-to-bacteria interactions within biofilms and in cell adherence and that EcpD forms a pilus tip adhesin.³¹

A preliminary computer-based search analysis of ecpA among the Enterobacteriaceae showed that a homolog of this gene is present in other species beyond E. coli, for example, Citrobacter rodentium, Shigella boydii, Klebsiella pneumoniae and Enterobacter (data not shown).²⁷ Further, this analysis showed that a homolog of the ecpRABCDE gen cluster is present in K. pneumoniae strains with the same genetic organization. Given the importance and established role of ECP in cell adherence and biofilm formation among the different E. coli pathotypes, we hypothesized that ECP could play a similar role for K. pneumoniae in its interplay with the host. Thus, the aim of the present study was to explore the biological significance of the prevalence and production of ECP among K. pneumoniae strains within the context of their interaction with host cells and their synergy with other known pili of the Klebsiella.

Results

The ecpA gene is highly prevalent among K. pneumoniae strains

We examined a collection of 69 K. pneumoniae strains including ATCC, normal flora, and clinical isolates for the presence of the ecpA gene encoding the major pilin subunit of the ECP by DNA amplification. PCR analysis using primers for ecpA flanking genes, ecpR and ecpB, showed that 58 (84%) of the strains were ecpA⁺ (Table 1). Previous studies have shown that genetic variations exist between ecpA genes of ETEC strains.³⁰ Thus, we employed primers specific to internal sequences of ecpA gene.³⁰ This approach rescued eight more ecpA⁺ strains that were initially PCR-negative. In sum, regardless of their origin, 66 (96%) of the 69 K. pneumoniae strains harbored the ecpA gene (Table 1).

Table 1. Analysis of pili genes and pili types in K. pneumoniae strains.

Total no. strains	Genotype*					Pili phenotype
n = 69	ecpRAB	ecpA	mrkA	mrkD	fimA†	ECP‡	MR/K‡	T1P§	Biofilm|| BF/ECP
No. positive for:	58	66	69	14	54	62	38	52	55/48
Percentage	84	96	100	20	78	90c	55	96	80/70

^*Based on PCR analysis. ^†fimA negative strains were fimH positive by PCR. ^‡Based on immunofluorescence assays. ^§Based on yeast agglutination assays. ^||No. of strains producing biofilms (BF) and ECP.

Analysis of the EcpA sequence from K. pneumoniae strains

The ecpA genes from three clinical isolates (Kpn 3, Kpn 10 and Kpn 07-216) were amplified and the purified amplicons subjected to nucleotide sequencing. The sequences obtained were deposited in the GenBank under accession numbers JN051492, JN051493 and JN051494. Alignment of predicted sequence protein of EcpA protein from EPEC E2348/69 and the EcpA protein from the three K. pneumoniae strains using ClustalW2 shows 3 conserved substitutions and 5 semi-conserved substitutions between them with 95% homology (Fig. S1; Table S1). The homology of EcpA among the three clinical K. pneumoniae strains was 99–100% and the homology with respect to the three previously sequenced K. pneumoniae strains was 97–100% (Fig. S2; Table S2).

Presence of mrkA, mrkD and fimA genes in K. pneumoniae strains

The PCR-based survey of fimA, fimH, mrkA and mrkD fimbrial genes in the K. pneumoniae collection showed that 69/69 (100%) of K. pneumoniae strains carried the mrkA gene while 54/69 (78%) carried fimA. We also examined these 15 fimA-negative strains for the presence of fimH and determined that all these isolates contained this gene. The mrkD adhesin gene was present in 14/69 (20%) strains (Table 1), indicating that not all strains harboring mrkA necessarily contain mrkD. These results are in agreement with previous reports in which the mrkD gene was found more conserved among K. oxytoca than in K. pneumoniae strains.²²

Ultrastructural analysis by electron microscopy of Klebsiella during infection of epithelial cells

Given the high number of K. pneumoniae strains positive for ecpA, we were interested in knowing if they produced ECP at all. Analysis of supernatants collected from HeLa infection assays of three selected K. pneumoniae ecpA⁺ strains (Kpn 3, Kpn Leo15 and Kpn 1) (Fig. 1A–C), and one K. pneumoniae ecpA⁻ strain (Kpn Leo40) (Fig. 1G and H), by transmission electron microscopy showed the presence of abundant peritrichous flexible or straight pili of varying length. To demonstrate whether the pili observed corresponded to ECP or MR/K we analyze the presence of ECP by immunogold EM using rabbit anti-ECP antibodies and secondary antibodies conjugated to colloidal gold particles. EPEC strain E2348/69, which produces ECP, was used as positive control and the derivative ΔecpA mutant was used as a negative control (data not shown).²⁸ Strains Kpn 3, Kpn Leo15 and Kpn 1 (Fig. 1D–F) were positive for ECP as shown by the gold labeling of surface-exposed filaments with anti-ECP antibodies, whereas Kpn Leo40 produced pili that did not correspond to ECP (Fig. 1G and H). Efforts to label MR/K fimbriae with the anti-MR/K antibodies by this immunogold labeling procedure were unsuccessful. Immunoblots of whole cell extracts of a subset of eight representative K. pneumoniae strains reacted with anti-ECP serum showed the 21-kDa EcpA pilin protein confirming the wide distribution and production of ECP (Fig. 2). Strains shown to be ecpA⁻ by PCR did not produce the EcpA pilin subunit.

Figure 1. Ultrastructural analysis of K. pneumoniae strains. Bacteria recovered from HeLa cells infection assays were negatively stained (A–C and G) or immuno-gold labeled (D–F, H and I) and visualized by transmission EM. (A and D) Kpn 3; (B and E) Kpn Leo15; (C and F) Kpn 1; (G) Kpn 40 showing pili, which did not reacted with anti-ECP antibodies (H). (I) K. oxytoca MXa showing ECP.

Figure 2. Detection of EcpA in selected K. pneumoniae strains. Immunobloting of whole cells extracts of Klebsiella strains treated with HCl and reacted with anti-ECP antibody. Purified ECP was used as a positive control. Strains Kpn 191–2, Kpn 292 and Kpn 11, which lack ecpA, were negative for EcpA.

ECP is produced by other Klebsiella species

We sought to determine if the ability to produce ECP expanded to other Klebsiella species. K. oxytoca strains MXa, 336 and 373 (MR/K⁺), which were pre-determined to be ecpA⁺ by PCR, were analyzed by immunogold labeling. The pili produced by the K. oxytoca strains also reacted positively with the anti-ECP antibodies (Fig. 1I and data not shown) indicating that both K. pneumoniae and K. oxytoca produce ECP. It is apparent that ECP is a property of Klebsiella species colonizing different biological and environmental niches.

ECP is produced by Klebsiella adhering to epithelial cells

In general, pili are considered virulence factors as they are required for cell adherence and biofilm formation. Thus, we hypothesized that the presence of ECP on bacteria infecting eukaryotic cells would conceivably signify a biologically relevant function. HeLa cells infected with three selected ECP⁺ and MR/K⁻ strains namely, Kpn 1, Kpn Leo15 or K. oxytoca 336, were examined by SEM and immuno-SEM using anti-ECP antibodies. These strains built large three-dimensional colonies containing large numbers of bacteria sitting on the surface of HeLa cells (Figs. 3A and B and 4A–C). This pattern of adherence is reminiscent of the localized adherence profile displayed by EPEC.^32,33 Observation at high magnification of these cell-associated bacterial clusters showed bacteria tethered to each other by a fibrillar network composed by copious amounts of long, flexible pili that extended several microns away from the bacteria (Fig. 3C and D). Employing back-scattered immuno-SEM, we confirmed that this meshwork of fibers contained ECP as evidenced by the presence of 30-nm gold particles on individual filaments (Fig. 3E and F). The SEM micrographs of K. oxytoca 336 show bacteria adhering tightly to each other onto the epithelial cell surface and to cellular filopodia through ECP (Fig. 4A–C). The distribution of the gold particles extending out from the bacteria strongly suggest that ECP function as anchoring devices such that individual pili filaments wrap around the thick cellular structures to facilitate colonization. In all these images strongly suggest that ECP must play an important role in cell adherence and to maintain strong physical contact between bacteria.

Figure 3. High-resolution SEM analysis of selected K. pneumoniae strains adhering to HeLa cells. (A, C and E) Kpn 1; (B, D and F) Kpn Leo15. (C and D) Images at high magnification showing numerous filaments connecting bacteria to each other in aggregates (arrows). (E and F) SIEM showing gold-labeled ECP on adhering bacteria.

Figure 4. High resolution SEM analysis of K. oxytoca 336 associated with the surface of HeLa cells. (A–C) Aggregates of strain 336 gold-labeled ECP connecting bacteria to each other and anchoring bacteria tightly to host cell filopodia structures (arrows). (D) Mock-infected HeLa cells were used as negative control.

Analysis of ECP and MR/K production by the K. pneumoniae collection

We sought to assess how many K. pneumoniae strains were actually able to assemble any of these two pili types during adhesion to cultured epithelial cells. Thus, we then examined the entire Klebsiella collection for surface assembly of ECP or MR/K on bacteria adhering to HeLa cells by immunofluorescence microscopy. The fluorescence profile obtained with anti-ECP antibodies showed long, fine, abundant peritrichous fibrillar structures (Fig. 5A, B and H) while the anti-MR/K antibodies detected MR/K as short knobs distributed peritrichously on the adhering bacteria (Fig. 5D, F and I). Isolate Kpn 3 was shown to co-produce ECP and MR/K (Fig. 5A, D and G) while Kpn Leo15 reacted with anti-ECP but not with anti-MR/K antibodies (Fig. 5B and E). Similarly, the pili displayed by Kpn Leo40 were labeled with anti-MR/K but not with anti-ECP antibodies (Fig. 5C, F and I). After studying the entire K. pneumoniae collection (69 strains) using this approach, it was determined that an overwhelming number of strains produced ECP [n = 62 (90%)] while 38 (55%) produced MR/K when adhering to HeLa cells, strongly suggesting that ECP production is a common feature among the Klebsiella strains (Table 1). Further, we found that 51% of the strains produced both ECP and MR/K, 39% produced ECP alone and 4% produced MR/K only (Table 2). Neither pilus type could be demonstrated in 6% of the strains tested. In addition, 48 (87%) out of the 55 mrkD⁻ strains were ECP⁺, seven (13%) mrkD⁻ were ECP⁻ and all of the mrkD⁺ strains produced ECP (Table 2).

Figure 5. Demonstration of ECP and/or MR/K on K. pneumoniae strains adhering to HeLa cells by immunofluorescence. (A, D and G) Kpn 3; (B, E and H) Kpn Leo15; (C, F and I) Kpn 40. ECP are shown in green and MR/K in red. Bacterial and cellular DNA was stained with DAPI (blue). Note the presence of both pili in Kpn 3. Images were taken at 60×.

Table 2. Co-production of MR/K and ECP in K. pneumoniae strains.

Total no. strains	No. of strains positive for:
n = 69	Fimbria and gene				Fimbriae production*
	mrkD+ (n = 14)		mrkD− (n = 55)		ECP+ (n = 62)		ECP− (n = 7)
	ECP+	ECP−	ECP+	ECP−	MRK/D+	MRK/D−	MRK/D+	MRK/D−
No. strains	14	0	48	7	35	27	3	4
Percentage	100	0	87	13	51	39	4	6

^*Production of fimbriae was determined by IFM during infections of HeLa cells.

Regarding T1P production, 52 (96%) out of the 54 fimA⁺ strains produced this pilus as determined by the yeast agglutination assay (Table 1). Further, the production of the pilin subunits MrkA, FimA and EcpA either alone or in combination, was demonstrated in a subset of K. pneumoniae strains by immunoblotting (Fig. 6). This experiment showed that there is a difference in the ability of these strains to produce one or more pili types.

Figure 6. Immunoblotting of whole cells extracts of selected K. pneumoniae strains. Bacterial extracts treated or not with HCl were reacted with the appropriate anti-pili antibodies to demonstrate the simultaneous production of 3 pili types.

Biofilm formation and ECP

Recent research suggests that ECP plays an important role in early stage biofilm development in meningitis-associated E. coli and attaching and effacing E. coli.^24,31 We set out to investigate the correlation between the ability of the K. pneumoniae strains to produce biofilms on a glass surface and to assemble ECP. We found that 55 (80%) of the 69 K. pneumoniae strains formed biofilms on glass coverslips (Table 1) and 48 (70%) of these biofilm-forming strains assembled ECP. As depicted in the micrographs shown in Figure 7, some strains showed abundant ECP while others showed moderate to no production of ECP. Of note, differences in ECP production were observed by bacteria within individual biofilms. For example, in Figure 7A, D, E and H around 50, 100, 20 and 5% of the bacteria forming biofilms produced ECP. These observations are reminiscent to those observed by pathogenic E. coli strains. It is reasonable to suggest that for a considerable number of K. pneumoniae and K. oxytoca strains, ECP is important for biofilm formation while other strains possess other biofilm-forming mechanisms.

Figure 7. Biofilm formation and ECP production in selected Klebsiella strains. (A) Kpn 1; (B) Kpn Leo15; (C) Kpn 25; (D) Kpn 3; (E) Kpn 13; (F) Klebsiella oxytoca 373; (G) Kpn 5; (H) Kpn JC. Biofilms obtained after 24 h of incubation at 37°C in DMEM were immuno-stained with anti-ECP antibodies (green) and bacteria were stained with propidium iodide (red). Images were taken at 60×.

Discussion

Although K. pneumoniae is a leading cause of Gram-negative nosocomial infections and is associated with a high mortality rate not much is known about its pathogenic scheme beyond the role of capsule.¹ In the established paradigm of K. pneumoniae pathogenesis, the anti-phagocytic polysaccharide capsule is the most distinguishing characteristic and most studied virulence factor.¹ Only a few other factors including LPS and pili have been proposed as virulence factors, but their contribution to host pathogenesis has not been examined in detail.¹ The ability of these bacteria to associate into communities in biofilms is central to their pathogenicity as they confer protection from bactericidal molecules present on host tissues. Key to the formation of biofilms is their property to produce a sticky capsule and several adhesive fimbrial (T1P, MR/K and KPF-28) and non-fimbrial (CF29K) adhesins and among which, the MR/K fimbriae have dogmatically been considered the most important adhesin of these bacteria.^8,11,12 Here we report for the first time that a large proportion (90%) of Klebsiella strains, including strains isolated from patients with hospital-acquired infections, produced ECP during epithelial cell infection in vitro and in biofilm-associated bacterial communities. The ultrastructural and immuno-detection analyses of Klebsiella strongly suggest that ECP serve to tether bacteria to each other leading to formation of defined microcolonies on cultured epithelial cells and to the progression of stable biofilms on inert surfaces. It is important to note that while 78% of strains were fimA, the remaining 22% carried fimH, demonstrating the wide distribution of the fim operon in all strains tested and that fimH is highly conserved among K. pneumoniae strains. These results are in agreement with our phenotypic data showing that 96% of the K. pneumoniae collection produced T1P. Similar frequencies of T1P (90%) were reported before in French isolates while 93% of the strains studied carried MR/K genes.³⁴ We were unable to detect T1P on Klebsiella strains adhering to HeLa cells using antibodies against E. coli T1P, probably due to intrinsic biochemical differences existing between Klebsiella and E. coli FimA proteins that presumably account for the differences in antibody recognition.³⁵ Because the infection assays were performed in the presence of D-mannose, which presumably blocks T1P-mediated adherence, it is assumed that T1P are not likely responsible for adherence to HeLa cells. Nevertheless, since the bacteria recovered from the infection assays are able to agglutinate yeast cells we cannot discard a role for T1P in cell adherence or bacteria-to-bacteria associations. Although all the strains contained the mrkA gene, only one half of them produced MR/K under the experimental conditions tested here. In agreement with past reports,^20,23 we found that most K. pneumoniae strains lack the mrkD gene, can still bind to host cells and form biofilms and produce MR/K and ECP.²² A high number of strains produced ECP in association with T1P, MR/K or both. Previous studies performed with French multi-drug resistant K. pneumoniae strains showed that 83% of them produced the KPF-28 fimbriae.¹¹ This fimbrial antigen was not sought here.

Similar to diarrheagenic E. coli strains, three adhesion patterns (aggregative, diffusive and localized) on cultured epithelial cells have been recognized for K. pneumoniae.^34,36 We also recognized these three adhesion patterns in the K. pneumoniae collection studied but could not find any association between the adherence phenotype and production of MR/K, T1P or ECP.

Biofilms are considered to play a role in the pathogenesis of numerous bacteria and are thought to play an important role in chronic infections due to the resistance they confer to bacteria to host immune responses and antimicrobials.³⁷ Eighty percent of the K. pneumoniae strains were able to form biofilms on plastic or glass surfaces and we detected ECP in a significant majority (70%) of the biofilm-producing strains. The level of ECP production within the biofilms varied from strain to strain. At the present time, we ignore if phase variation occurs in ECP expression. Some strains produced thick biofilms independently of ECP or MR/K indicating that other surface factors may exist in Klebsiella that aid in biofilm development. Lehti et al., found that Mat fimbriae, a homolog of ECP, are required for biofilm formation in meningitis-causing E. coli strains.²⁵ To provide genetic evidence of the involvement of ECP in bacterial adherence and biofilm formation, we made several efforts to obtain an ecpA mutant in K. pneumoniae strain, using the lambda Red mutagenesis system but we were unsuccessful. One of the practical problems when working with Klebsiella strains is the multi-drug resistance of these organisms, which limits the number of resistance cassettes that can be employed in the laboratory to mutagenize bacterial genes. While further research is needed to clarify the precise role of ECP in the pathogenesis of Klebsiella, and how the multiple pili are synchronized during infections (e.g., through the analysis of individual pili mutants), this study provides compelling evidence that ECP may be an important factor for cell adherence, formation of biofilms and colonization of several niches, particularly for strains lacking the MrkD adhesin or the entire MR/K pilus. Moreover, this study highlights the heterogeneity of K. pneumoniae strains and their potential to produce multiple pili types which conceivably could play a significant synergistic role during host colonization.

Materials and Methods

Bacterial strains and growth conditions

The collection of 69 K. pneumoniae strains used in this study included two American Type Culture Collection (ATCC) reference strains, four isolates from healthy individuals, 49 clinical isolates obtained from patients in Mexico with pneumonia or septicemia and 14 isolates from patients with nosocomial infections in the United States. The clinical isolates were obtained from blood, pleural fluid, stool and urine specimens and were resistant to multiple antibiotics including β-lactams, fluoroquinolones and carbapenems. Bacteria were routinely grown overnight in Luria-Bertani (LB) broth or in Dulbecco modified Eagle’s medium (DMEM) (Invitrogen) at 26°C for induction of ECP. EPEC strain E2348/69 and its derivative isogenic ΔecpA were used as positive and negative controls of ECP production, respectively.²⁸ E2348/69 ΔfimE (T1P positive strain) and ΔfimBE (T1P negative strain) mutants were used as positive and negative controls of T1P production, respectively (Saldana Z. et al., unpublished).

Gene amplification

Amplification of ecpA, mrkA, mrkD, fimA and fimH genes in K. pneumoniae strains was done by PCR using GoTaq Green Master Mix (Promega) containing 1 mM of each specific primers (listed in Table 3) as previously described.^29,30 In addition, the ecpA genes from three clinical isolates (Kpn 3, Kpn 10 and Kpn 07-216) were amplified using Platinum Taq DNA polymerase High Fidelity (Invitrogen) and primers for ecpA flanking genes, ecpR and ecpB (Table 3), and the products were subjected to nucleotide sequencing at the Interdisciplinary Center for Biotechnology Research (ICBR) Nucleotide Sequencing Facility, University of Florida (UF). The reaction was run for 35 cycles with a melting temperature of 94°C, an annealing temperature of 62°C and an elongation temperature of 72°C.

Table 3. Primers used in this study.

Primer	Sequence	Target gene	Product size
G510	5′-CCTATGTAATTAATGGCAGGTTT-3′	ecpRAB	1025 bp
G511	3′-GCTGTTCATAAAGGATGAAATATC-5′
G569	5′-GCAACAGCCAAAAAAGACACC-3′	ecpA	477 bp
G570	3′-CCAGGTCGCGTCGAACTG-5′
G595	5′-CTGACGCTTTTTATTGGCTTAATGGCGC-3′	mrkD	757 bp
G596	3′-GCAGAATTTCCGGTCTTTTCGTTTAGTAG-5′
G593	5′-CGGTAAAGTTACCGACGTATCTTGTACTG-3′	mrkA	498 bp
G594	3′-GCTGTTAACCACACCGGTGGTAAC-5′
fimA- F	5′-CGGACGGTACGCTGTATTTT-3′	fimA	500 bp
fimA- R	3′-GCTTCGGCGTTGTCTTTATC-5′
fimH- F	5′-TGGTGGTCGACCTCTCCACGCAGATTTTTTGCC-3′	fimH	576 pb
fimH- R	3′-TCAGCTGAACGCCTATCCCCTGCGCCGGCGAGGCGG-5′

Transmission electron microscopy and immunogold labeling

ECP was visualized on bacteria cultured overnight at 26°C in DMEM or from supernatants collected from HeLa cell-infection assays (37°C), after negative staining with 1% phosphotungstic acid (pH 7.2) using a Hitachi 7600 transmission electron microscope.²⁶ Immuno-electron gold labeling was performed using highly absorbed and specific rabbit anti-ECP described in previous studies^26,28,30 or anti-MR/K (kind gift of Dr. Steven Clegg) antibodies (diluted 1:10 in 1% bovine serum albumin in PBS) and goat anti-rabbit IgG conjugated to 10 nm-gold particles (diluted 1:10) with 1 h incubations each and negatively stained as before.

Adherence and immunofluorescence assays

Ten microliters of overnight bacterial cultures were incubated for 6 h with 80% confluent HeLa cell monolayers at 37°C under an atmosphere of 5% CO₂ as previously described.²⁸ After thorough washing with PBS, the cells were fixed with 2% formalin in PBS for immunofluorescence in order to visualize ECP on bacteria adhering to HeLa cells using rabbit anti-ECP or anti-MR/K antibodies diluted (1:5,000) in PBS-10% horse serum and secondary goat anti-rabbit IgG conjugated with Alexa-Fluor 488.²⁸ For double-labeling experiments, chicken anti-ECP and rabbit anti-MR/K antibodies were co-incubated with infected cells followed by incubation with goat anti-rabbit IgG (for MR/K) conjugated with Alexa-Fluor 532 and goat anti-chicken IgG (for ECP) conjugated with Alexa-Fluor 488 (diluted 1:5,000) for 1 h at room temperature. After washing with PBS, 1:10,000 dilution of DAPI in PBS was used for DNA staining. Coverslips were mounted on glass slides with 0.1% p-phenylenediamine mounting medium. Samples were visualized under a Zeiss AXIO ImagerA1 microscope and photographed with a QImaging Retiga 2000R Fast 1394 Camera.²⁸

Scanning electron microscopy (SEM)

After infection, HeLa cells were fixed overnight at 4°C with Trump’s Fixative (Electron Microscopy Sciences) and processed for SEM by the ICBR Electron Microscopy Facility at UF as previously described.³⁸ The samples were examined with a field-emission scanning electron microscope (S-4000, Hitachi High Technologies America, Inc.). For immuno-SEM, infected HeLa cells on coverslips were fixed with 4% paraformaldehyde, 1% glutaraldehyde in 1X PBS, pH 7.24 and processed as described above. Fixed cells were washed in PBS and incubated with blocking solution [1.5% BSA, 0.5% cold water fish skin gelatin, 0.01% Tween-20 in PBS (pH 7.2) for 1 h]. After overnight incubation with anti-ECP polyclonal antibody diluted (1:100 in PBS) and goat anti-rabbit IgG conjugated to 30-nm gold particles (Ted Pella) was added at room temperature for 1 h. Immunogold stained cells were post fixed and examined as described above.

Biofilm assay

K. pneumoniae strains were cultivated statically in DMEM at 37°C and 100 µL of these cultures were used to inoculate 24-well plates containing glass coverslips and 1 mL of DMEM. After 24 h of incubation, the biofilms produced were fixed with 2% formalin, stained with 1% propidium iodide, and the pili immuno-stained with anti-ECP and the appropriate fluorochromogenic conjugate and then visualized under light epifluorescence using an AxioImager 1.0 Zeiss microscope.

SDS-PAGE and western blotting

Bacterial pellets obtained by centrifugation of overnight DMEM-grown cultures were boiled for 5 min in acidic water (pH 1.8). Sample denaturing buffer (4X) was added and the pH was then neutralized with NaOH before SDS-PAGE in 16% polyacrylamide gels. Western blotting was performed using rabbit anti-pili (ECP, T1P, MR/K) antibodies (1:5,000) and an anti-rabbit IgG horseradish-peroxidase conjugate (1:20,000) as previously described.²⁶ The blots were developed with ECL Plus Western Blotting Detection System (GE Healthcare).

Yeast agglutination assay

Bacteria present in the supernatants from the HeLa cell infections described above were collected, pelleted and adjusted with PBS to an equal density. A colony of Saccharomyces cerevisiae strain 4450 grown for 2 d on YPD agar was mixed with one drop of the bacterial suspension on a glass slide and rocked for 30–60 sec until appearance of aggregates visible to the naked eye. To determine if agglutination was due to the presence of T1P, samples were also processed in the presence of 3% D-mannose.

Disclosure of Potential Conflicts of Interest

No potential conflicts of interest were disclosed.