Expression of a Clostridium perfringens Type IV Pilin by Neisseria gonorrhoeae Mediates Adherence to Muscle Cells

Abstract

Clostridium perfringens is an anaerobic, Gram-positive bacterium that causes a range of diseases in humans, including lethal gas gangrene. We have recently shown that strains of C. perfringens move across the surface of agar plates by a unique type IV pilus (TFP)-mediated social motility that had not been previously described. Based on sequence homology to pilins in Gram-negative bacteria, C. perfringens appears to have two pilin subunits, PilA1 and PilA2. Structural prediction analysis indicated PilA1 is similar to the pseudopilin found in Klebsiella oxytoca, while PilA2 is more similar to true pilins found in the Gram-negative pathogens Pseudomonas aeruginosa and Neisseria gonorrhoeae. Strains of N. gonorrhoeae that were genetically deficient in the native pilin, PilE, but supplemented with inducible expression of PilA1 and PilA2 of C. perfringens were constructed. Genetic competence, wild-type twitching motility, and attachment to human urogenital epithelial cells were not restored by expression of either pilin. However, attachment to mouse and rat myoblast (muscle) cell lines was observed with the N. gonorrhoeae strain expressing PilA2. Significantly, wild-type C. perfringens cells adhered to mouse myoblasts under anaerobic conditions, and adherence was 10-fold lower in a pilT mutant that lacked functional TFP. These findings implicate C. perfringens TFP in the ability of C. perfringens to adhere to and move along muscle fibers in vivo, which may provide a therapeutic approach to limiting this rapidly spreading and highly lethal infection.

INTRODUCTION

Clostridium perfringens is a Gram-positive anaerobic pathogen that causes gas gangrene and food poisoning (32). Recently, we have shown that C. perfringens possesses a unique gliding motility that is mediated by type IV pili (TFP) (39). TFP in other bacteria are implicated in attachment to and invasion of host cells, microcolony and biofilm formation, attachment to abiotic surfaces, twitching motility, bacteriophage susceptibility, and genetic competence (1, 13, 17, 23, 28, 39, 40). TFP are long, flexible filaments that extend from the bacterial cell surface and are homopolymers composed of thousands of copies of a single protein, pilin (4, 15). Pilins are synthesized as precursor proteins, with an N-terminal leader sequence that is cleaved by a bifunctional endopeptidase (PilD), usually at a phenylalanine residue immediately preceded by a glycine residue; the phenylalanine residue is also methylated by the same endopeptidase (30, 38).

The crystal structures of several type IV pilins from Gram-negative bacteria have been solved (3, 7, 8, 12, 29). Pilin structures reveal a conserved core comprised of an extended N-terminal α-helix and a globular domain. The pilins have a ladle shape due to the protrusion of the N-terminal helix from the globular domain (8). Two regions flank the conserved structural core and are exposed on the surfaces of the pilus filaments: the αβ-loop and the D-region, delineated by two cysteines linking the C-terminal segment to the β-sheet via a disulfide bond (6). These regions are believed to perform many of the pilus-associated functions (10).

Two gene products, PilA1 and PilA2, were identified as putative pilins that serve as structural subunits of TFP in C. perfringens (36, 39). In a previous report (39), we described results obtained with the FUGUE protein structure prediction program (35), which we used to compare the C. perfringens PilA proteins to those in the database (available on the FUGUE server at http://www-cryst.bioc.cam.ac.uk/∼fugue/). PilA1 proteins had only a single match to the pseudopilin PulG from Klebsiella oxytoca. However, the PilA2 proteins had matches to both PulG and pilins from Pseudomonas aeruginosa and Neisseria gonorrhoeae (39). Since pseudopilins have similar folds to pilins but lack the D-loop domain, the similarity of the PilA2 protein to both the pseudopilins and the pilins is logical, since pseudopilins comprise a subset of the larger pilin structures. The conserved cysteine pair seen in the Gram-negative type IV pilins is not present in PilA2, but the C-terminal segment is still present and is predicted to form a loop similar to the D-region of PilE (39).

The N-terminal α-helices comprise the pilin polymerization domain, based on cryo-electron microscopy (cryo-EM) reconstruction of the N. gonorrhoeae type IV pilus (8). The N-terminal α-helices are arranged in a helical array inside the filament core, making extensive contacts with one another to provide mechanical strength. The globular domains line the outer core of the pilus, exposing the αβ-loop and D-regions (6). The C. perfringens pilus PilA2, along with all type IV pili, likely shares the same helical architecture as other pilins, based on conserved sequence and structural and biological functions associated with all TFP systems analyzed thus far.

The structure of the N-terminal truncated PulG pseudopilin from Klebsiella oxytoca (18), which is similar to the predicted PilA1 structure, shows a similar fold in the N-terminal domain and globular head region as that seen in true pilins but lacks the D-loop containing the hypervariable region between the disulfide bond. The three-dimensional structure allowed a model for pseudopilus stacking in a filament to be constructed, where 17 monomers comprised four helical turns in a left-handed filament (18). The pseudopilin filaments are predicted to be much shorter than those in TFP, which corresponds to the observation that pseudopilins do not extend past the outer membrane in Gram-negative bacteria.

Additional evidence supporting the assignment of PilA1 as a pseudopilin and PilA2 as a true pilin comes from a phylogenetic analysis of the conserved region (i.e., the first 65 to 69 amino acids) of pilin and pseudopilin proteins (Fig. 1A; see also Fig. S1 in the supplemental material). The analysis was carried out using previously described methods (31). The bootstrap analysis showed pilins and pseudopilins fall into two distinct families of proteins and that C. perfringens PilA1 is in the pseudopilin family and PilA2 in the pilin family (Fig. 1A). In addition, the bootstrap values for PilA1 and PilA2 have a top score, suggesting they are both derived from an independent line of descent within the pilin or pseudopilin group (Fig. 1A). This supports our previous hypothesis that clostridia TFP are ancient and likely the ancestral forms from which all Eubacteria TFP systems have evolved (39).

Fig. 1.

(A) Phylogenetic tree showing relatedness of the conserved N-terminal domains of pilin and pseudopilin proteins. Common colors denote evolutionarily related pilins. Bootstrap values are shown on individual branches. They were calculated from 100 replicates using the neighbor method in the PAUP 4.0.4a software (42). Note that PilA1 and PilA2 sequences branched with a top score, suggesting their line of descent within the pseudopilin or pilin group is independent. The alignments used to generate this tree are shown in Fig. S1 of the supplemental material. (B) Alignment of pilin proteins PilA1 and PilA2 from C. perfringens strain 13 with PilE from N. gonorrhoeae strain MS11 using the Clustal W 2.0 program (19). Residues with an asterisk are identical in all three sequences, those with a colon are similar in all three sequences, and those with a period are similar or identical in 2 of 3 sequences. The arrow indicates the site of prepilin peptidase cleavage (known in PilE and proposed in PilA1 and PilA2).

With this structural information in hand, we wanted to assign a specific function to each pilin. Because C. perfringens pili are short (200 to 300 nm) and difficult to distinguish from one another on intact cells (39), we decided to determine if PilA1 and PilA2 can complement a deficiency in PilE, the native pilin, in N. gonorrhoeae strain MS11. N. gonorrhoeae was chosen because (i) the pilus assembly apparatus has been well characterized, (ii) heterologous expression of Pseudomonas aeruginosa PilA (43) and Francisella tularensis PilA (33) in N. gonorrhoeae results in formation of intact pili, and (iii) N. gonorrhoeae exists as single cells or diplococci, which makes visual observations significantly easier in comparison to C. perfringens, in which motile bacteria are found in long chains of cells with an end-to-end conformation (39). Our results showed that PilA2 from C. perfringens was assembled into pili but PilA1 was not. While neither pilin from C. perfringens complemented the absence of native pilin for genetic competence or adherence to human urogenital cells, PilA2 pili enabled the N. gonorrhoeae strain expressing this protein to adhere to myoblasts (muscle cells) derived from rodent cell lines. C. perfringens, under anaerobic conditions, was also able to adhere to mouse myoblasts in a TFP-dependent manner, suggesting the results seen with N. gonorrhoeae are relevant to the host species itself.

MATERIALS AND METHODS

Bacterial strains and growth conditions.

The bacterial strains used in this study are listed in Table 1. Escherichia coli DH10B and DH5α were used for cloning steps and were grown in Luria-Bertani (LB) broth, supplemented with 15 g/liter agar (Difco) for solid medium. Chloramphenical at 20 μg/ml and 100 μg/ml kanamycin were added when necessary for selection. C. perfringens strains 13, SM125 (pilT mutant), and SM125(pSM271) (complemented pilT mutant) were grown anaerobically as previously described (39).

Table 1.

Bacterial strains, plasmids, and oligonucleotides used in the study

Strain, plasmid, or oligonucleotide	Relevant characteristic(s) or sequence	Source or reference
E. coli DH10B	F−mcrA Δ(mrr-hsdRMS-mcrBC) φ80dlacZΔM15 lacX74 deoR recA1 araD139 (ara leu)7697 galU galK I-rpsL endA1 nupG	Gibco/BRL
C. perfringens strains
Strain 13	Gangrene-associated strain	20
SM125	pilT mutant	39
SM125(pSM271)	pilT mutant complemented with wild-type pilT	39
N. gonorrhoeae strains
MS11	Wild type, ΔpilE2	34
MS11-307	ΔpilE1::erm ΔpilE2	26
MS11-310	ΔpilE1::kan	2
KR106	ΔpilE1::erm ΔpilE2, C. perfringens pilA2+	This study
KR107	ΔpilE1::erm ΔpilE2, C. perfringens pilA1+	This study
Plasmids
pKH35	E. coli origin of replication, lacIq, OPOPlac, three Neisseria uptake sequences, chloramphenicol resistance	11
pKR106	pKH35 with C. perfringens pilA2 and E. coli ribosome binding site	This study
pKR107	pKH35 with C. perfringens pilA1 and E. coli ribosome binding site	This study
pKR112	PGem-T Easy with pilT from N. gonorrhoeae MS11	This study
pKR113	pKR112, Kanr inserted into SacII site in pilT	This study
Oligonucleotides
OKR011	5′-CTTTAAAAAATAAAGCTTAGGAGGCATAAATGTTATTACTGAAAGC-3′
OKR013	5′-CATATTAAGCTTAGGAGGAAAACCAATGAATACAAAAAAAC-3′
OKR014	5′-CACCTTATTATTTTACTAATTCTAGAACTTTTAATACTATTGATTATTTC-3′
OKR019	5′-GCCCCTATGAAATCTAGATAATTAAAGTATATTTATACTATAGACACC-3′
OKR063	5′-CGAAACCATATGCTAAACAGAGCCGCA-3′
OKR064	5′-GCCTTGCCATGGCGCC-3′
OKR065	5′-CCGCGGCCGTCTGAACTCAAAATC-3′
OKR066	5′-CCGCGGTTGATGAGAGCTTTGTTGTAG-3′

N. gonorrhoeae strains were derivatives of MS11A (34). N. gonorrhoeae strain MS11-307 (ΔpilE1::erm ΔpilE2) served as the nonrevertible nonpiliated control (2, 26). In liquid culture, all N. gonorrhoeae strains were grown at 37°C and 5% CO₂ in GC broth supplemented with 0.042% sodium bicarbonate and Kellogg's supplements I and II (16). Solid medium was prepared by supplementing GC medium base (Difco) with 1.25 g/liter agar and Kellogg's supplements. Erythromycin at 3 μg/ml and 10 μg/ml chloramphenicol were added as needed. Piliation and Opa⁻ phenotypes were assessed microscopically.

Plasmid construction.

Plasmids and oligonucleotides used in this study are summarized in Table 1. C. perfringens pilA1 and pilA2 genes were amplified by PCR using primer pairs OKR011/OKR019 and OKR013/OKR014, respectively. Amplification yielded a 484-bp pilA1 gene and a 609-bp pilA2 gene. PCR products were digested with HindIII and XbaI and ligated into similarly digested pKH35 (11). This placed the pilA1 and pilA2 genes immediately downstream of the lac promoter, yielding plasmids pKR107 and pKR106. Clones were verified by DNA sequencing.

Generation of N. gonorrhoeae strains expressing C. perfringens PilA1 and PilA2.

A derivative of the N. gonorrhoeae strain MS11 (MS11C3; ΔpilE2) was transformed with pKR106 and pKR107. Resulting transformants were then transformed with genomic DNA from MS11-307 (ΔpilE1::Erm) (26) to yield KR106 and KR107, nonpiliated mutants that harbor pilA2 and pilA1, respectively.

Induction and detection of pilA1 and pilA2 expression.

N. gonorrhoeae strains were grown on GC agar (GCA) plates as described above with isopropyl-β-d-thiogalactopyranoside (IPTG) added to a concentration of 0.5 mM. Colonies grown overnight were swabbed from plates, resuspended in SDS-PAGE loading buffer (100 mM dithiothreitol, 50 mM Tris-HCl, 2% SDS, 0.1% bromophenol blue, and 10% glycerol), and boiled for 20 min. Samples were centrifuged for 8 s at 12,500 rpm in a microcentrifuge, and 15 μl of the supernatant was subjected to SDS-PAGE. Western blot assays using antibodies to PilA1 and PilA2 were carried out as previously described (39).

Growth curves.

One colony each from strains KR106 and KR107 was inoculated into 20 ml GCB supplemented with 0.042% sodium bicarbonate and incubated at 37°C with shaking at 300 rpm. Overnight cultures were diluted 1/20 with fresh medium and incubated at 37°C with shaking. IPTG (0.5 mM) was added to one 20-ml culture of KR106 and one KR107 culture, and 1.0 mM IPTG was added to a separate culture of each strain. One milliliter of each culture was removed every hour, serially diluted with fresh medium, and plated on GCA. Plates were incubated at 37°C overnight, and colonies were counted to calculate CFU/ml.

Attempted pilT mutagenesis.

The pilT gene was amplified from MS11 chromosomal DNA with oligonucleotides OKR063 and OKR064, using Phusion high-fidelity DNA polymerase (New England BioLabs). The PCR product was cloned into pGEM-Teasy (Promega), resulting in plasmid pKR112. The kanamycin resistance gene was amplified by PCR from chromosomal DNA of an N. gonorrhoeae strain carrying the EZ::Tn5<KAN-2> transposon using primers OKR065 and OKR066 and was then cloned into the pilT gene to produce plasmid pKR113. The linear fragment consisting of the pilT gene interrupted by the kanamycin resistance cassette was used for electroporation with recipient N. gonorrhoeae strains MS11 and KR106, as previously described (9).

Immunofluorescence.

N. gonorrhoeae strains KR106 and KR107, grown for 24 h with and without IPTG (1 mM), were swabbed and resuspended in GC broth to an optical density at 600 nm of 0.1. Ten microliters of each culture was placed on a glass coverslip and incubated at 37°C for 1.5 h and then fixed by adding 700 μl of 2.5% paraformaldehyde in Dulbecco's phosphate-buffered saline (DPBS) for 10 min. Coverslips were then washed in 2% bovine serum albumin (BSA) in DPBS (BSA/DPBS) and blocked in 10% goat serum in DPBS at 37°C for 30 min. The coverslips were washed twice in BSA/DPBS and then incubated with affinity-purified antibodies (39) against PilA1 or PilA2 in BSA/DPBS at 37°C for 30 min. Coverslips were then washed twice in 2% BSA in DPBS and incubated with 1:100 goat anti-rabbit serum conjugated to AlexaFluor 594 (Molecular Probes). Coverslips were subjected to a final wash with 2% BSA in DPBS, and images were captured using an Olympus IX81 upright microscope linked to a Hamamatsu model C4742 charge-coupled-device (CCD) camera. Slidebook 4.1 imaging software (Intelligent Imaging Innovations) was used to collect images.

Transformation assays.

N. gonorrhoeae strains were swabbed from plates grown for 18 h into 1 ml GC broth supplemented with 5 mM MgCl₂ and Kellogg's supplements I and II (GCB+Mg+supps). The concentration of bacteria was determined spectrophotometrically and adjusted to 10⁸ CFU/ml. Twenty-microliter aliquots of these cultures and 200 ng of genomic DNA from N. gonorrhoeae strain MS11-310 (ΔpilE1::Kan) were added to 180 μl GCB+Mg+supps and incubated at 37°C and 5% CO₂ for 1 h to allow DNA uptake. The samples were then diluted 10-fold with GCB+Mg+supps and incubated for 4 h at 37°C with 5% CO₂. Serial dilutions of the transformation mixes were plated on nonselective and selective (100 μg/ml kanamycin) GCA and incubated 48 h at 37°C and 5% CO₂. Transformation efficiency was calculated by dividing the number of Kan^r transformants by the total number of bacteria present at the end of the experiment.

N. gonorrhoeae adherence assays.

For A431 (human epidermal carcinoma cell line) and MDCK (canine kidney cell line), 10⁴ cells per well were seeded into a 24-well plate in Dulbecco's modified Eagle's medium (DMEM) for A431 cells or Eagle's minimum essential medium (EMEM) for MDCK cells and incubated at 37°C and 5% CO₂ until approximately 90% confluent. GC strains were swabbed from plates grown for 24 h and then used to inoculate 1 ml prewarmed supplemented GCB. KR106 and KR107 were swabbed from plates without IPTG as well as plates containing 0.5 mM IPTG and diluted to a concentration of 4 × 10⁶ CFU/ml in DMEM for adherence to A431 cells or EMEM for adherence to MDCK cells. Cell culture media were supplemented with 10% heat-treated fetal bovine serum (FBS) and Kellogg's supplement II. Spent cell culture medium was removed from the seeded eukaryotic cells, and each well was inoculated with 250 μl of the appropriate bacterial strain. Six wells were inoculated per N. gonorrhoeae strain and incubated for 3 h at 37°C and 5% CO₂. Three wells per strain were designated “total bacteria” wells and three were designated “adherent bacteria” wells. Total bacteria were calculated as follows: 250 μl of supernatant (total liquid volume) was removed from the well and set aside. Then, 750 μl PBS with 0.25% EDTA was added to each well and incubated at room temperature for 10 min, after which cells that remained attached were removed by vigorous pipetting. This 750 μl of bacteria and detached cells was added to the original 250 μl removed from the well, and dilutions were made in PBS followed by plating on GCA. Colonies were counted after 48 h of incubation at 37°C and 5% CO₂. For adherent bacteria, the supernatant was aspirated and each well was washed five times with PBS. Then, 1 ml of PBS containing 0.25% EDTA was added to each well and allowed to sit for 10 min. Cells that remained attached were lifted by vigorous pipetting. Following mixing with a vortexer for 20 s, dilutions were prepared in PBS and plated on GCA supplemented with Kellogg's supplements I and II. Colonies were counted after 48 h of incubation at 37°C and 5% CO₂. The percent adherence of each strain was determined by dividing the number of adherent bacteria by the number of total bacteria per well and then multiplying by 100. Induced expression of PilA1 and PilA2 was confirmed by Western blotting, as described above.

For C2C12 myoblasts (46) and L6 myoblasts (22), cells were seeded into 48-well plates containing DMEM with 10% FBS (DMEM/FBS) and incubated at 37°C and 5% CO₂ until approximately 90% confluence was observed. N. gonorrhoeae strains were swabbed from plates incubated for 24 h and used to inoculate 1 ml of prewarmed supplemented GCB. KR106 and KR107 were swabbed from plates without IPTG as well as plates containing 0.5 mM IPTG. Each strain was diluted in DMEM supplemented with 10% heat-treated FBS and Kellogg's supplement II to a concentration of 4 × 10⁸ CFU/ml. The spent cell culture medium was removed from the seeded myoblasts, and each well was inoculated with 250 μl of the appropriate bacterial strain. A total of 11 wells were inoculated per GC strain. These were incubated at 37°C and 5% CO₂ for 3 h. Total bacteria and adherent bacteria were calculated as described above, with the exception that myoblasts were lifted from wells by scraping in three directions with a pipette tip and then pipetting vigorously up and down. Expression of PilA1 and PilA2 from strains KR107 and KR106, respectively, was confirmed by Western blotting on induced cultures used in these assays.

C. perfringens adherence assays.

Adherence of C. perfringens strains to C2C12 cells took place in a Coy anaerobic chamber with an atmosphere of 85% N₂, 10% CO₂, and 5% H₂. C2C12 cells were grown to confluence (∼2 days) in 0.5 ml DMEM/FBS in 48-well tissue culture plates and then placed in a 37°C incubator inside the anaerobic chamber. C. perfringens strains, grown overnight on PGY medium with agar (24) under anaerobic conditions, were removed from the anaerobic chamber, scraped off the plates, and suspended in 1 ml DPBS. Bacteria were pelleted in a centrifuge and suspended in DPBS, and the suspensions were diluted in DPBS to give ∼2 × 10⁷ CFU/ml. After the C2C12 cells were incubated under anaerobic conditions for 2 h, 5 μl of the bacterial suspension (∼1 × 10⁵ CFU) was added to each well and incubated anaerobically at 37°C for 75 min. The plates were then removed from the anaerobic chamber, and each well was washed 3 times with 0.5 ml aerobic DPBS to remove unattached bacteria. After the final wash, 0.5 ml of distilled water was added to the wells to lyse the myoblasts. The cells and bacteria were scraped off the bottom of the well, placed in a microcentrifuge tube, and subjected to vortex mixing for 20 s. The bacteria in the sample were then quantified by serial dilution and plating on PGY medium. Attachment assays for each strain were performed on quintuplicate samples from two separate experiments (10 samples total). C2C12 cell viability before and after incubation under anaerobic conditions was tested using the trypan blue dye exclusion assay, as previously described (27).

Video microscopy.

All video was obtained using an Olympus IX81 upright microscope linked to a Hamamatsu model C4742 CCD camera. Slidebook 4.1 (Intelligent Imaging Innovations) imaging software was used to compile motility videos. To determine if strains KR106 and KR107 exhibited twitching motility, colonies of each strain were removed from plates incubated overnight and resuspended in GCB that was prewarmed to 37°C and supplemented with Kellogg's supplements and 0.042% sodium bicarbonate. Ten-microliter aliquots of the bacterial suspensions were added to microscope slides, covered with coverslips, and immediately inverted. Slides were maintained at 37°C during microscopy.

To examine the interactions of C. perfringens with C2C12 cells under anaerobic conditions, C2C12 cells were first grown aerobically in DMEM/FBS in 50-ml tissue culture flasks until reaching ∼50% confluence. They were then transferred to a 37°C incubator inside a Coy anaerobic chamber that contained an atmosphere of 85% N₂, 10% CO₂, and 5% H₂. After 2 h under anaerobic conditions, a 1:1 multiplicity of infection (MOI) of C. perfringens cells was added to the flask. The cap on the flask was tightened to prevent oxygen from getting inside, and the flask was placed on a heated stage set at 37°C for video microscopy.

Electron microscopy.

For field emission scanning electron microscopy (FE-SEM) to visualize pili, N. gonorrhoeae strains were grown overnight on GC agar and Kellogg's supplements and then spread on 0.2 μm GTTP filters (Millipore) on the surface of fresh GC agar plates with 1 mM IPTG and incubated at 37°C in a CO₂ incubator for 2 h. After 2 h of growth, the filters were lifted from the plates and mounted on top of aluminum sample holders embedded in dry ice, which quickly froze the samples. The filters were then sputter coated with gold to a thickness of 1 nm, and pili were imaged using a Helios 600 NanoLab SEM (FEI). For SEM of N. gonorrhoeae strains attached to C2C12 myoblasts, attachment was obtained by seeding C2C12 myoblasts on 13-mm round glass coverslips layered on glass beads in a 12-well tissue culture plate and allowed to grow to approximately 90% confluence. N. gonorrhoeae strains were added and incubated as described above for attachment assays, and samples on glass coverslips were transferred to 6-well plates containing 2 ml of 2.5% glutaraldehyde in 0.1 M sodium cacodylate (pH 7.2) and fixed overnight at 4°C. Samples were postfixed for 30 min in 1% osmium tetroxide in cacodylate buffer. The coverslips were then washed in deionized water, dehydrated in ethanol, and critical point dried using a Ladd critical point dryer. Dried samples were mounted onto sample holders, lightly sputter coated with gold, and examined with a Phillips 505 scanning electron microscope.

RESULTS

Pilin alignments.

Alignments of PilA1 and PilA2 from C. perfringens strain 13 with PilE from N. gonorrhoeae strain MS11 are shown in Fig. 1B. Overall sequence homology between PilA1/PilA2 and PilE was low, except in the N-terminal region (Fig. 1B). Even less sequence similarity was seen when PilA1 and PilA2 were compared to the minor pilins PilX and PilV of N. gonorrhoeae (data not shown). Both PilA1 and PilA2 from C. perfringens possess conserved N-terminal leader sequences at the site of prepilin peptidase (PilD) cleavage (Fig. 1B), which suggests they are likely to be processed into mature pilins.

Expression of PilA1 and PilA2 in N. gonorrhoeae.

N. gonorrhoeae strains KR107 and KR106 were constructed to express C. perfringens PilA1 and PilA2, respectively, under an IPTG-inducible promoter in a pilE mutant background using previously described methods (11). Addition of 0.5 mM or 1 mM IPTG had no effect on the growth rate of either strain (see Fig. S2 in the supplemental material). In the presence of IPTG, strain KR107 (PilA1 expression) had a colony pilin phenotype identical to the PilE mutant strain, MS11-307, while KR106 (PilA2 expression) had a colony pilin phenotype somewhere between MS11-307 and MS11 (PilE⁺). Pilin proteins were detected by Western blotting using affinity-purified anti-PilA1 and anti-PilA2 antibodies on whole-cell lysates prepared from GC strains KR107 and KR106, respectively, after induction with IPTG. Neither protein was detected in the absence of IPTG (Fig. 2). PilA1 expressed in N. gonorrhoeae appeared to be 14 kDa, which runs slightly larger than the predicted size of the processed protein, at 12.6 kDa. However, PilA2 expressed in N. gonorrhoeae is 23 kDa, larger than the predicted size of 19.5 kDa. To determine if PilA2 was posttranslationally modified in N. gonorrhoeae, PilA2-containing cell extracts were treated with O-glycosidase, N-glycosidase, and sialidase, in series and individually and with both Antarctic phosphatase and alkaline phosphatase. There was no change in the apparent mass of PilA2 after any of these treatments (data not shown). These results suggest PilA2 is either covalently modified with an unknown moiety or migrates somewhat slower than predicted in SDS-PAGE gels.

Fig. 2.

Western blot analysis showing heterologous expression of C. perfringens pilins in N. gonorrhoeae. Whole-cell lysates of colonies swabbed from plates were subjected to SDS-PAGE followed by immunoblotting with anti-PilA1 and anti-PilA2 antibodies. The locations of molecular mass standards are indicated on the left. +, cultured in the presence of IPTG; -, cultured in the absence of IPTG.

Immunofluorescence using affinity-purified antibodies against PilA1 and PilA2 showed increased expression of each of these proteins on the surface of N. gonorrhoeae cells after induction by IPTG (Fig. 3). The fluorescence appeared in specific regions on the surface of the cells (Fig. 3), which may correspond to the location of pili on the surface of the bacteria, but pili themselves, if present, would not be resolved using light microscopy due to their small size.

Fig. 3.

Immunofluorescence (red color) shows increased expression of PilA1 and PilA2 on the surface of N. gonorrhoeae strains KR106 and KR107 in the presence of IPTG. Bars, 2 μm. DIC, differential interference contrast microscopy.

Pili were visible by FE-SEM on strain MS11, the N. gonorrhoeae wild-type strain, grown on GTTP filters (Fig. 4A). Despite examining thousands of cells of each type, no pili could be seen with the pilE mutant strain, MS11-307, or the strain expressing C. perfringens pilA1, KR107 (data not shown). Strain KR106, expressing the pilA2 gene in the presence of IPTG, produced numerous pili associated with ∼10 to 15% of the cells (Fig. 4B and data not shown). Measurements of the length of pili that could be clearly defined as being attached to bacteria showed the PilA2 fibers were similar in length to PilE fibers, 1.06 ± 0.35 μm and 1.08 ± 0.29 μm, respectively (means ± standard deviation). Many PilA2 fibers were visible on the surface of filters in regions lacking bacteria (Fig. 4C), suggesting they had detached during growth or while processing the filters for electron microscopy. A high magnification image of an isolated PilA2 fiber (Fig. 4D) showed the fibers are ∼7 nm in diameter, similar to that seen with fibers on the surface of C. perfringens (39).

Fig. 4.

Field emission-scanning electron microscopy of N. gonorrhoeae strains cultured on 0.2-μm membrane filters. (A) Wild-type N. gonorrhoeae strain MS11. (B) N. gonorrhoeae strain KR106, a pilE mutant expressing C. perfringens PilA2. (C) Surface of a filter near strain KR106 cells; unattached pili can be seen on the surface of the filter. (D) High-magnification image of an isolated PilA2 pilus from strain KR106. Regions inside boxes are shown at higher magnification in the insets in panels A and B. Bar, 1 μm (A to C) or 100 nm (D).

Complementation of pilin functions.

Because TFP are necessary for genetic transformation in N. gonorrhoeae, strains KR106 and KR107 were assayed for their abilities to restore transformation in the absence of pili. Wild-type strain MS11 transformed at a frequency of 2 × 10⁻³ transformants per CFU using genomic DNA from a kanamycin-resistant N. gonorrhoeae strain. However, no transformants were detected in the same experiment with the pilE mutant (MS11-307) or N. gonorrhoeae expressing PilA1 or PilA2 from C. perfringens (data not shown).

TFP are also needed for twitching motility in N. gonorrhoeae. Video microscopy was used to determine if PilA1- and PilA2-expressing strains exhibited twitching motility. The cells in Video S1 of the supplemental material clearly demonstrate the characteristic twitching motility of wild-type N. gonorrhoeae on a glass coverslip. N. gonorrhoeae strain MS11-307, a pilE mutant, did not show twitching motility, and the cells of this strain were motionless (see Video S2 in the supplemental material). Similarly, strain KR107 bacteria could be seen binding to both the coverslip (cells in focus cells) and the glass slide (cells out of focus) but exhibited very little motion in the presence of IPTG (see Video S3 in the supplemental material), indicating that expression of PilA1 from C. perfringens does not restore twitching motility in the absence of native PilE. N. gonorrhoeae strain KR106 possesses a twitching motility phenotype that resembles neither the twitching motility of wild-type N. gonorrhoeae nor the motionless phenotype of the pilE mutant. In the presence of IPTG, some KR106 cells appeared to adhere to the glass coverslip and spin in place (see Video S4 in the supplemental material). We hypothesize that the diplococci can extend a pilus composed of C. perfringens PilA2, but that the pilus cannot easily be retracted, so the cell remains tethered to the glass coverslip and rotates due to Brownian motion. While only a fraction of the cells in any given video frame exhibited this phenotype (see Video S4), this fraction roughly corresponded to the fraction of cells possessing pili, as seen by FE-SEM (Fig. 4B).

Adherence to mammalian cells of N. gonorrhoeae expressing PilA1 and PilA2.

N. gonorrhoeae strains expressing PilA1 and PilA2 were tested for their ability to mediate adherence to human epithelial cells, which in N. gonorrhoeae is partially dependent on TFP (21). Wild-type N. gonorrhoeae adhered to A431 cells (a vulval carcinoma cell line) at an average frequency of 33% (Fig. 5A). The pilE mutant strain and strain KR106 in the absence or presence of IPTG did not adhere appreciably to this cell line (Fig. 5A). We observed nonspecific binding of strain KR107 bacteria, but none of the other strains tested, to the plastic tissue culture plates in the absence of mammalian cells (data not shown), and therefore, adherence data for this strain are not included.

Fig. 5.

Adherence of N. gonorrhoeae strains expressing C. perfringens pili. Adherence of N. gonorrhoeae strains to A431 cells (a human vulval carcinoma cell line) (A), C2C12 cells (a mouse myoblast cell line) (B), L6 cells (a rat myoblast cell line) (C), mouse primary fibroblasts (D), and MDCK cells (a canine kidney cell line) (E). Values shown are means and standard deviations of at least 3 independent samples. Numerical values above each bar represent the P values, calculated using a two-tailed unpaired t test with Welch correction. Note differences in scales used in each panel.

Since C. perfringens strain 13 is an etiologic agent of gas gangrene, which is an infection of muscle tissue, we rationalized that C. perfringens was likely to encounter muscle cells and connective tissue cells during a gangrene infection. Therefore, an N. gonorrhoeae strain expressing PilA2 was tested for its ability to adhere to two different myoblast cell lines, C2C12 and L6, derived from mice and rats, respectively. Wild-type N. gonorrhoeae adhered to C2C12 cells at a frequency of 0.0005%, which was not statistically different from the adherence of the nonpiliated mutant, at 0.02% (Fig. 5B). KR106 adhered to C2C12 cells at a rate of 0.15% in the presence of IPTG, but adherence dropped to 0.02% in the absence of IPTG (Fig. 5B). This was significantly higher than that of wild-type N. gonorrhoeae, the nonpiliated mutant, and uninduced KR106 (P < 0.001).

Surprisingly, wild-type N. gonorrhoeae adhered to L6 myoblasts (rat) at a rate of 2.13% (Fig. 5C). The pilE mutant adhered at a rate of 0.02%, which was 100-fold lower and statistically significant (P < 0.0066). PilA2 expression in strain KR106 resulted in adherence to L6 cells at a rate of 0.17%, which was significantly higher than observed for strain KR106 without IPTG (P < 0.0256), which adhered at a rate of 0.012%, a 14-fold difference. We also tested cells derived from a type of connective tissue, primary mouse fibroblasts, and found that both strains MS11 and KR106 with IPTG adhered significantly better than the corresponding strains, MS11-307 and KR106 without IPTG (Fig. 5D), suggesting either pilus protein could mediate adherence.

In order to assess the adherence of each strain to a cell type in which it was expected that neither N. gonorrhoeae native pili nor C. perfringens pili would bind, assays were performed with MDCK (canine kidney) cells (Fig. 5E). As projected, adherence to this cell line was less than 0.003% for all strains tested, and no strain adhered at a significantly higher frequency than any other.

Visual confirmation of adherence to mouse muscle cells (C2C12) was obtained by SEM of C2C12 cells cultured on glass coverslips. Cells of strain KR106 appeared as either microcolonies (Fig. 6A) or isolated diplococci (Fig. 6B). Strain KR107 can be seen binding to both the smooth surface of the coverslip and C2C12 cells in SEM images (Fig. 6C), confirming our observation that PilA1 expression caused N. gonorrhoeae to adhere to inert surfaces, such as plastic plates and glass coverslips, in the absence of mammalian cells.

Fig. 6.

SEM of N. gonorrhoeae strains KR106 and KR107 adhering to mouse muscle cells. C2C12 cells were seeded on glass coverslips and incubated with N. gonorrhoeae strain KR106 (A and B) or N. gonorrhoeae strain KR107 (C), all with IPTG added. Bar, 1 μm.

Adherence of C. perfringens to C2C12 cells.

Since PilA2 expression allowed N. gonorrhoeae to adhere to rodent myoblasts, we performed adherence assays with live C. perfringens strain 13 bacteria, the source of the pilA1 and pilA2 genes used in this study. Under anaerobic conditions in liquid medium (DMEM/FBS) >95% of the C2C12 myoblasts were viable for 12 h (data not shown), as measured using trypan blue exclusion. Initial attachment assays using cells with fresh DMEM/FBS showed undetectable adherence of C. perfringens bacteria to C2C12 cells under anaerobic conditions (data not shown). Video microscopy of the C. perfringens-C2C12 interactions under anaerobic conditions showed the bacteria were not lined up in the characteristic end-to-end fashion seen in motile C. perfringens cells (39) but rather floated in the tissue culture medium mostly as individual cells (see Video S5 in the supplemental material). We theorized that glucose, present in fresh DMEM at 4.5 g/liter, might be inhibiting the expression of motility genes, as previously reported (25, 39). Therefore, we carried out the same assay using C2C12 cells that had been grown for 2 days in DMEM to deplete the glucose before placement in an anaerobic chamber. Under these conditions, C. perfringens strain 13 adhered at a relatively high level of 3.8% (Fig. 7). A strain 13 mutant with an inactivated pilT gene, SM125, adhered ∼10-fold less efficiently (Fig. 7). When strain SM125 was complemented with a multicopy plasmid carrying an intact pilT gene, adherence increased ∼2-fold, but not to the same level as the wild-type strain (Fig. 7). Video microscopy of the C2C12 cells and strain 13 bacteria under low-glucose conditions showed the bacteria now formed chains and aggregates that appeared to be connected in a head-to-tail manner (see Video S6 in the supplemental material), similar to what was seen on agar surfaces (39). In contrast, the pilT mutant strain, while composed of somewhat longer cells than those seen with strain 13, did not form long chains or aggregates (see Video S7 in the supplemental material).

Fig. 7.

Adherence of C. perfringens strains to C2C12 cells under anaerobic conditions. Values shown represent the means and standard deviations of 10 independent samples. Numerical values above each bar represent the P values, calculated using a two-tailed unpaired t test with Welch correction.

DISCUSSION

Structural prediction programs provided evidence that C. perfringens PilA1 is a pseudopilus while PilA2 is a true pilin (39). In an effort to obtain information about the actual functions of these pilins in C. perfringens, we constructed strains of N. gonorrhoeae that expressed these proteins. The pilA1 gene was expressed and the protein made (Fig. 2), but no pili could be seen on the surface of the cells via FE-SEM (data not shown). In contrast, the PilA1 protein could be found on the surface of intact N. gonorrhoeae by using immunofluorescence (Fig. 3). We interpret these results as indicating the PilA1 protein is somehow exposed on the surface to allow antibodies to bind but is not present as intact pili. This may be due to its predicted pseudopilin-like structure, which might not allow the N. gonorrhoeae pilus assembly apparatus to polymerize it into pili. The observation that PilA1 expression mediated attachment to artificial surfaces such as plastic and glass (Fig. 6) may correlate with the presence of this protein on the surface of the bacteria.

Pseudopili are often associated with type II secretion systems (14). High-level heterologous expression of the Klebsiella pneumoniae pseudopilus PulG in E. coli resulted in the formation of bundled extracellular pili (41). However, unlike E. coli, N. gonorrhoeae does not possess a type II secretion system of its own, which may explain why the results from our experiments differed from those obtained by Vignon et al. (41). Altogether, these results indicate the function of PilA1 in C. perfringens remains to be determined.

The N. gonorrhoeae strain expressing pilA2, KR106, formed pili that resembled those visible by FE-SEM on the surface of C. perfringens in width (5 to 7 nm [39]). However, the pili visible on the surface of C. perfringens are 0.2 to 0.3 μm long (39), while PilA2 made by N. gonorrhoeae are ∼1 μm long (Fig. 4). Because we were unable to determine what protein comprised the surface pili on C. perfringens via immunogold bead labeling due to their short length and high density (39), it is possible that these pili fibers were not comprised of PilA2 but of some other protein. Alternatively, it may be that pilus polymerization terminates more quickly in C. perfringens than N. gonorrhoeae, which would account for the difference in lengths seen with these species.

While able to form pili, N. gonorrhoeae strain KR106 could not complement the natural transformation and twitching motility phenotypes seen with wild-type N. gonorrhoeae. However, strain KR106 did provide a unique motility phenotype. In contrast to wild-type N. gonorrhoeae cells, which appeared to crawl across the glass coverslip (see Video S1 in the supplemental material), a fraction of the KR106 cells examined appeared to adhere to the coverslip and rotate in place (see Video S4), possibly due to Brownian motion of bacteria suspended at the end of a tether. Because competence and normal twitching motility are not restored in KR106, and both of these functions require the action of the retraction ATPase PilT (44, 45), we hypothesize that PilA2 is sufficiently compatible with the TFP apparatus in N. gonorrhoeae to be assembled into pili but that PilA2 is unable to interact with PilT, so pili cannot be retracted. This hypothesis was tested by attempting to create a pilT mutation in strain KR106 by introducing recombinant DNA by electroporation (see Materials and Methods), since strain KR106 is not competent. Despite numerous attempts, no mutants were obtained using this method, perhaps due to the sensitivity of strain KR106 to the high voltages required for electroporation (data not shown).

Expression of PilA2 mediated attachment of N. gonorrhoeae to mouse and rat myoblasts as well as mouse fibroblasts but not to human vulval carcinoma cells or canine kidney cells (Fig. 5). Interestingly, this pattern of attachment corresponds to cell types that C. perfringens would likely encounter in a gangrene infection (i.e., muscles and connective tissue), but not to other cell types. To our knowledge, this is the first specific adherence factor that has been identified in C. perfringens.

We also demonstrated that TFP mediate adherence of C. perfringens to C2C12 cells if the conditions are optimal (Fig. 7). These conditions include anaerobiosis and low glucose levels, both of which are likely to be found in the environment favorable for gas gangrene to develop and spread. The pilT mutant derivative of strain 13, SM125, showed a 10-fold decrease in adherence, suggesting 90% of the adherence is dependent on PilT-related functions. The levels of PilA1 and PilA2 on the surface of strain SM125 were lower than those seen with strain 13, which differs from what has been observed in most other TFP systems in Gram-negative bacteria, where pilT mutants are hyperpiliated (39). Recently, a pilT mutant of the Gram-negative pathogen Francisella tularensis was shown to lack surface pili (5), indicating this phenotype exists in Gram-negative bacteria as well.

Lack of PilA1 and/or PilA2 on the surface may account for the decreased adherence we observed with strain SM125. We do not know why the plasmid carrying a wild-type copy of the pilT gene failed to fully complement the adherence phenotype but did allow for motility on agar plates (39). One possibility is that multiple copies of the pilT gene produced abnormally high levels of the PilT protein, a retraction ATPase, thereby shifting the equilibrium of pilin polymerization more to the retracted state with fewer pili on the surface of the cells.

A distinct characteristic of gas gangrene is the remarkable extent to which muscle tissue is degraded (37). Adherence to muscle fibers may allow C. perfringens to move along a muscle bundle and not be detached by the physical action of muscle contraction and relaxation. Also, by mediating both adherence and motility, TFP may permit the bacteria to surround muscle fibers in three dimensions. This would allow them to break down the fibers more effectively with degradative enzymes than if they were confined to a single side of a muscle bundle.

N. gonorrhoeae adhered to rat muscle cells and mouse fibrobalsts in a PilE-dependent manner (Fig. 5B), although adherence to the rat cells was much more efficient (Fig. 5C and D). N. gonorrhoeae is considered to be a human-specific pathogen and does not naturally infect rodents, and so, although of little pathogenic significance, we postulate that there is a gratuitous receptor present on L6 cells but not on C2C12 cells with which native N. gonorrhoeae pili can interact. This unknown receptor is likely not the same one to which PilA2 binds, since PilA2 mediated attachment to both L6 and C2C12 cells (Fig. 5B and C).

Our finding that N. gonorrhoeae can polymerize PilA2 from C. perfringens into an intact pilus is a testament to the extreme level of conservation TFP systems have retained throughout evolutionary history. Given the apparent independent line of descent of PilA2 from Gram-negative pilins (Fig. 1A) and the proposal that the clostridia are likely the ancestral form of Eubacteria TFP (39), polymerization of PilA2 into intact pili by N. gonorrhoeae suggests this TFP assembly system has maintained the same basic structural properties since the earliest stages of bacterial evolution.