Email updates

Keep up to date with the latest news and content from BMC Microbiology and BioMed Central.

Open Access Highly Accessed Research article

Expression of the Flp proteins by Haemophilus ducreyi is necessary for virulence in human volunteers

Diane M Janowicz1*, Sean A Cooney1, Jessica Walsh2, Beth Baker3, Barry P Katz1, Kate R Fortney1, Beth W Zwickl1, Sheila Ellinger1 and Robert S Munson34

Author affiliations

1 Department of Medicine, Indiana University, School of Medicine, Indianapolis, IN 46202, USA

2 Department of Microbiology and Immunology, Indiana University, School of Medicine, Indianapolis, IN 46202, USA

3 The Research Institute at Nationwide Children's Hospital, The Ohio State University, Columbus, Ohio 43205, USA

4 Department of Pediatrics, The Ohio State University, Columbus, Ohio 43205, USA

For all author emails, please log on.

Citation and License

BMC Microbiology 2011, 11:208  doi:10.1186/1471-2180-11-208

The electronic version of this article is the complete one and can be found online at: http://www.biomedcentral.com/1471-2180/11/208


Received:1 June 2011
Accepted:22 September 2011
Published:22 September 2011

© 2011 Janowicz et al; licensee BioMed Central Ltd.

This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Abstract

Background

Haemophilus ducreyi, the causative agent of the sexually transmitted disease chancroid, contains a flp (fimbria like protein) operon that encodes proteins predicted to contribute to adherence and pathogenesis. H. ducreyi mutants that lack expression of Flp1 and Flp2 or TadA, which has homology to NTPases of type IV secretion systems, have decreased abilities to attach to and form microcolonies on human foreskin fibroblasts (HFF). A tadA mutant is attenuated in its ability to cause disease in human volunteers and in the temperature dependent rabbit model, but a flp1flp2 mutant is virulent in rabbits. Whether a flp deletion mutant would cause disease in humans is not clear.

Results

We constructed 35000HPΔflp1-3, a deletion mutant that lacks expression of all three Flp proteins but has an intact tad secretion system. 35000HPΔflp1-3 was impaired in its ability to form microcolonies and to attach to HFF in vitro when compared to its parent (35000HP). Complementation of the mutant with flp1-3 in trans restored the parental phenotype. To test whether expression of Flp1-3 was necessary for virulence in humans, ten healthy adult volunteers were experimentally infected with a fixed dose of 35000HP (ranging from 54 to 67 CFU) on one arm and three doses of 35000HPΔflp1-3 (ranging from 63 to 961 CFU) on the other arm. The overall papule formation rate for the parent was 80% (95% confidence interval, CI, 55.2%-99.9%) and for the mutant was 70.0% (95% CI, 50.5%-89.5%) (P = 0.52). Mutant papules were significantly smaller (mean, 11.2 mm2) than were parent papules (21.8 mm2) 24 h after inoculation (P = 0.018). The overall pustule formation rates were 46.7% (95% CI 23.7-69.7%) at 30 parent sites and 6.7% (95% CI, 0.1-19.1%) at 30 mutant sites (P = 0.001).

Conclusion

These data suggest that production and secretion of the Flp proteins contributes to microcolony formation and attachment to HFF cells in vitro. Expression of flp1-3 is also necessary for H. ducreyi to initiate disease and progress to pustule formation in humans. Future studies will focus on how Flp proteins contribute to microcolony formation and attachment in vivo.

Background

The tad (tight adherence) locus is present in many Gram-positive and Gram-negative bacteria as well as the Archaea and likely represents an ancient subtype of type IV secretion systems that was horizontally transferred to many bacterial species early in the course of evolution. The tad genes are located on a mobile genomic island coined "the widespread colonization island" by Figurski and coworkers [1]. The functions of the tad locus gene products have been best described for Aggregatibacter (formerly Actinobacillus) actinomycetemcomitans, where they are essential for adherence, biofilm formation, and pathogenesis. The Tad proteins are predicted to form a macromolecular transport system for the assembly and secretion of fimbria or pili, which mediate adherence of the organism to different surfaces [2].

Haemophilus ducreyi is a gram-negative coccobacillus that causes the sexually transmitted genital ulcer disease chancroid, which is a public health concern because it increases the risk of transmission and acquisition of human immunodeficiency virus-1. H. ducreyi contains a 12.8 kb flp (fimbria like protein) operon with 15 genes (flp1-flp2-flp3-orfBC-rcpAB-orfD-tadABCDEFG) that encode products with homology to proteins encoded by the tad locus in A. actinomycetemcomitans [3,4]. H. ducreyi mutants that lack expression of either Flp1 and Flp2, which encode fimbria like proteins, or tadA, which has homology to NTPases of type IV secretion systems, have decreased abilities to attach to human foreskin fibroblasts (HFF) and to form microcolonies on HFF [4,5]. A mutant deficient in the expression of tadA was significantly but partially attenuated in the temperature dependent rabbit model of experimental chancroid, while a mutant lacking flp1 and flp2 expression was fully virulent in this model [4,5]. Thus, whether Flp-Tad-mediated adherence and/or microcolony formation are critical factors in the virulence of H. ducreyi is unclear [6].

In experimental and natural infection in humans, H. ducreyi forms aggregates, the first step in microcolony formation, and colocalizes with polymorphonuclear leukocytes and macrophages, which fail to ingest the organism. In human inoculation experiments, a tadA mutant is highly attenuated for virulence; whether the observed attenuation is due to the lack of secretion of the Flp proteins or other unidentified effectors by the tad locus is unclear [5]. Given the discrepancy in virulence between the tadA mutant and the flp1flp2 mutant in the temperature dependent rabbit model [5], here we constructed and characterized a flp1-3 deletion mutant. We tested the flp1-3 mutant for its ability to cause disease in human volunteers and its ability to form microcolonies and adhere to human fibroblasts. Our data indicate that expression of Flps is required for virulence and that Flp-Tad mediated adherence correlates with the virulence of H. ducreyi in humans. To our knowledge, this study is the first to provide definitive proof that expression of the Flp proteins is required for the virulence of a bacterial pathogen in humans.

Results

Construction and characterization of 35000HPΔflp1-3

An unmarked, in frame deletion mutant of the flp1, flp2, and flp3 genes was constructed in 35000HP using recombineering technology and designated 35000HPΔflp1-3 [7,8]. Sequence analysis of 35000HPΔflp1-3 confirmed that flp1, flp2 and flp3 had been replaced by a short ORF that consisted of the upstream region of flp1, the start codon of flp1, 81 bp encoding a scar peptide, and the last 21 bp of flp3, including its stop codon. By qRT-PCR, the expression levels of tadA and tadG, two genes downstream of flp3, were similar in 35000HPΔflp1-3 compared to 35000HP (data not shown), suggesting that the remainder of the flp operon was normally transcribed. 35000HP and 35000HPΔflp1-3 demonstrated identical growth rates in broth (data not shown). The LOS profiles and OMP patterns as analyzed by SDS-PAGE were similar for the mutant and the parent (data not shown).

Human inoculation experiments

To determine whether the Flp proteins play a role in pathogenesis, 35000HPΔflp1-3 was compared with 35000HP for virulence using a mutant parent comparison trial in the human model of infection. Ten healthy adults (six males, four females; 5 Caucasian, 5 black; age range 32 to 59; mean age ± standard deviation, 48 ± 9 years) volunteered for the study. Three subjects (volunteers 333, 334, and 335) were inoculated in the first iteration, three subjects (volunteers 336, 337, and 338) in the second iteration, one subject (volunteer 341) in the third iteration and three subjects (volunteers 342, 343, and 344) in the fourth iteration. In the first iteration, each subject was infected with a fixed EDD (61 CFU) of 35000HP at three sites on one arm and varying EDDs (63, 125, and 249 CFU) of 35000HPΔflp1-3 on the other arm (Table 1). Since the ability of the mutant to form pustules seemed impaired (Table 1), we increased the dose of the mutant in subsequent iterations, as per protocol. In the second iteration, volunteers were inoculated with 65 CFU of the parent and 66, 131, and 262 CFU of the mutant (Table 1). In the third iteration, the volunteer was infected with 67 CFU of 35000HP and 240, 480, and 961 CFU of the mutant (Table 1). In the fourth iteration, volunteers were inoculated with 54 CFU of the parent and 104, 208 and 415 CFU of the mutant (Table 1).

Table 1. Response to inoculation of live H.ducreyi strains

The overall papule formation rate for the parent was 80% (95% confidence interval, CI, 55.2%-99.9%) at 30 sites and for the mutant was 70.0% (95% CI, 50.5%-89.5%) at 30 sites (P = 0.52). Mutant papules were significantly smaller (mean, 11.2 mm2) than were parent papules (21.8 mm2) 24 h after inoculation (P = 0.018). The overall pustule formation rates were 46.7% (95% CI 23.7-69.7%) at 30 parent sites and 6.7% (95% CI, 0.1-19.1%) at 30 mutant sites (P = 0.001). Mutant pustules formed at only two sites in one volunteer. These results indicate that expression of one or more of the flp1, flp2, and flp3 genes in the context of the intact secretion/assembly complex is necessary for H. ducreyi to initiate disease and progress to pustule formation in humans.

H. ducreyi was recovered intermittently from surface cultures. Of the 30 sites that were inoculated with the parent, 11 (36.7%) yielded at least one positive surface culture, while 3 of 30 mutant sites (10.0%) yielded a positive surface culture (P = 0.019). All colonies recovered from sites inoculated with the parent (n = 626) or the mutant (n = 39) and colonies from the parent (n = 142) and mutant (n = 143) inocula were tested for the presence of flp1-flp2-flp3 and fgbA sequences by colony hybridization. The fgbA probe hybridized to all the colonies, while the flp1-2-3 probe hybridized only to the colonies obtained from the parent inoculated sites or the parent inocula. Thus, there was no cross contamination of mutant and parent sites during the course of the trial.

Deletion of flp1flp2flp3 affects attachment of 35000HP to HFF cells and microcolony formation

To confirm the phenotype of 35000HPΔflp1-3, we complemented 35000HPΔflp1-3 in trans by electroporating it with plasmid pJW1, which encodes the Flp1, Flp2 and Flp3. flp1, flp2, and flp3 encode proteins with predicted molecular weights of 9.3 kDa, 8.9 kDa, and 9.9 kDa, respectively. Western blot analysis with a polyclonal sera that binds to Flp1 and Flp2 confirmed that 35000HPΔflp1-3(pLSSK) lacked the ability to express the Flp1 and Flp2 proteins (Figure 1, lane 2) compared to 35000HP(pLSSK) (Figure 1, lane 1). Complementation of 35000HPΔflp1-3 with plasmid pJW1 resulted in restoration of the expression of the Flp1 and Flp2 proteins as determined by Western blot (Figure 1, lane 3).

thumbnailFigure 1. Western Blot analysis of Flp1 and Flp2 expression by wild type, mutant, and complemented H. ducreyi strains. Whole-cell lysates were probed with polyclonal rabbit Flp1 antiserum as the primary antibody. Lanes: 1, wild-type 35000HP(pLSSK); 2, 35000HPΔflp1-3(pLSSK); 3, 35000HPΔflp1-3(pJW1). Molecular markers are shown on the left.

35000HP(pLSSK), 35000HPΔflp1-3(pLSSK), and 35000HPΔflp1-3(pJW1) were also tested for their abilities to bind confluent HFF monolayers. 35000HPΔflp1-3(pLSSK) significantly attached to HFF cells at lower levels (geomean ± standard deviation, 26.0% ± 15.0%) than did 35000HP(pLSSK) (100% ± 29.0%) (P = 0.018) (Figure 2). 35000HPΔflp1-3(pJW1) adhered to HFF cells (92.0% ± 18.0%) at significantly higher levels than 35000HPΔflp1-3(pLSSK) (P = 0.010) and at similar levels as 35000HP(pLSSK) (P = 0.32) (Figure 2).

thumbnailFigure 2. Quantitative measurement of the binding of wild type, mutant, and complemented H. ducreyi strains to HFF cells. Assays were performed as described in Materials and Methods. The data represented are a composite of five separate experiments. Bars: 1, wild-type 35000HP(pLSSK); 2, 35000HPΔflp1-3(pLSSK); 3, 35000HPΔflp1-3(pJW1).

35000HP(pLSSK), 35000HPΔflp1-3(pLSSK), and 35000HPΔflp1-3(pJW1) were also compared for their abilities to form microcolonies after 24 h incubation with confluent HFF monolayers. 35000HP formed numerous, densely populated microcolonies on the surfaces of HFF cells [4] (Figure 3A). 35000HPΔflp1-3(pLSSK) formed sparse and very small microcolonies (Figure 3B) when compared to 35000HP; the complemented mutant demonstrated a restored phenotype similar to 35000HP(pLSSK) (Figure 3C). Thus, complementation of the mutant restored the parental phenotypes.

thumbnailFigure 3. Microcolony formation by (A) wild type 35000HP(pLSSK), (B) flp1-3 mutant 35000HPΔflp1-3(pLSSK), and (C) complemented flp1-3 mutant 35000HPΔflp1-3 (pJW1). Magnification ×400.

Discussion

For this study, we focused on whether the expression of the Flp proteins was necessary for virulence of H. ducreyi. We constructed an unmarked, in frame deletion mutant lacking the flp1flp2flp3 genes in 35000HP using a recombineering strategy [8,9] and found that 35000HPΔflp1-3 was significantly impaired in its ability to cause disease in the human model of infection. flp1-3 joins hgbA, dsrA, ncaA, lspA1-lspA2, pal, tadA sapBC and cpxA as the ninth gene required for full virulence by H. ducreyi in the human inoculation experiments [9,10] (unpublished observations).

The flp-tad gene cluster is constitutively transcribed as a single polycistronic operon in vitro [4]. Relative to its expression during in vitro growth, tadA transcripts are enriched in experimental pustules, suggesting that the flp-tad operon is upregulated in vivo [11]. CpxRA is the only obvious intact two-component signal transduction system contained in H. ducreyi. Transcription of flp1-3 and several other major virulence determinants are negatively regulated by conditions that favor phosphorylation of CpxR [9,12,13]. Purified recombinant CpxR interacts with the promoter regions of the flp operon in electrophoretic mobility shift assays [13]. Deletion of cpxA leads to loss of CpxA phosphatase activity, activates CpxR, and cripples the ability of H. ducreyi to infect humans [9]. In contrast, a cpxR deletion mutant has no effect on or upregulates the expression of virulence determinants and is fully virulent in human volunteers [13]. Taken together, the data suggest that the flp-tad operon may be upregulated in vivo due to downregulation of CpxRA.

The human inoculation experiments are limited in that we are precluded by several regulatory bodies from testing trans-complemented mutants in humans. However, complementation of 35000HPΔflp1-3 in trans restored the ability of the mutant to form microcolonies and bind to HFF cells, suggesting that the phenotype of the mutant is due to the deletion of the flp genes. In the human inoculation experiments, we use 35000HP to examine the role of virulence factors in H. ducreyi pathogenesis. There are two classes of H. ducreyi strains, which express different immunotypes and proteomes [14,15]. Although we were able to amplify flp1-3 alleles from six class I and three class II strains (data not shown), attempts to sequence the amplicons were unsuccessful, so we do not know if there is a difference in the flp genes in the class I and class II strains. 35000HP is a class I strain; whether the Flp proteins play a role in the virulence of class II strains is unknown.

We previously reported that a tadA mutant is attenuated for pustule formation in the human challenge model [5]. However, the tadA mutant, but not a flp1flp2 double mutant, is attenuated in the rabbit model of chancroid [4,5]. Nika et al previously reported that both the flp1flp2 mutant and the tadA mutant demonstrate decreased abilities to attach to HFF cells and form fewer microcolonies on HFF cells [4]. These data suggested that microcolony formation by itself is not a virulence factor for H. ducreyi. Although H. ducreyi does not appear to co-localize with fibroblasts in experimental or natural chancroid [16,17], our data indicate that adherence to HFF cells in vitro correlates with the virulence of H. ducreyi in humans. Similarly, both flp1 and tadA mutants fail to colonize or cause disease in a rat infection model with A. actinomycetemcomitans [1], and both flp1 mutant and tadD mutants of Pasteurella multocida are attenuated in the septicemic mouse model [18,19].

In A. actinomycetemcomitans, Flp pili are assembled as bundles of long fibers in which Flp1 is the major structural component [3,20]. However, there is no evidence that the Flp proteins are assembled into a pilus-like structure in H. ducreyi [4]. Several bacterial species including A. actinomycetemcomitans have two flp genes [2]. H. ducreyi contains three flp genes, which have between 50-80% similarity to one another [4]. Deletion of flp1 and flp2 results in decreased adherence of H. ducreyi to HFF cells and subsequent microcolony formation [4]; the function of Flp3 is unclear.

In vitro, H. ducreyi forms microcolonies, a key step in biofilm formation. In vivo, H. ducreyi forms aggregates and colocalizes with macrophages, PMNs, collagen and fibrin [16,17]. H. ducreyi contains a luxS homologue that has autoinducer (AI-2) activity in a Vibrio harveyi-based reporter system, and a luxS mutant is partially attenuated for virulence in human volunteers [21]. Taken together, these data suggest that the formation of microcolonies, aggregates and quorum sensing mechanisms may be important for H. ducreyi pathogenesis. Whether the Flp proteins contribute to this process by mediating attachment to host cells or initiating microcolony formation in the skin remains a subject for future investigation.

Conclusions

We have constructed an unmarked, in frame deletion mutant lacking the flp1flp2flp3 genes in H. ducreyi strain 35000HP. The deletion mutant, 35000HPΔflp1-3, has an intact tad secretion system. Our data show that production and secretion of the Flp proteins contributes to microcolony formation and attachment of 35000HP to HFF cells in vitro. Complementation of the mutant with flp1-3 in trans restored the parental phenotype. Additionally, expression of Flp1-3 is necessary for H. ducreyi to initiate disease and progress to pustule formation in humans. Future studies will focus on how Flp proteins contribute to microcolony formation and attachment in vivo.

Methods

Bacteria and culture conditions

35000HP is a human-passaged (HP) variant of strain 35000 and has been reported previously [22]. H. ducreyi strains were grown on chocolate agar plates supplemented with 1% IsoVitaleX at 33°C in 5% CO2. For the human inoculation experiments, H. ducreyi was grown in a protease peptone broth-based medium supplemented with 50 μg of hemin per ml, 1% IsoVitaleX and 5% heat-inactivated fetal calf serum (FCS) as described [23] or in a Columbia broth based medium with 2.5% heat-inactivated FCS for other experiments. When appropriate, the media were supplemented with chloramphenicol, spectinomycin, or kanamycin at 0.3 μg/ml, 200 μg/ml, or 20 μg/ml, respectively, to maintain plasmids or select for chromosomal integration of antibiotic resistance cassettes. E. coli strains were grown in Luria-Bertani broth or plates at 37°C, except for strain DY380, which was grown or maintained at 32°C and induced to express λ recombinase at 42°C.

Construction and characterization of a flp1-3 mutant of strain 35000HP

An unmarked, in frame deletion mutant of the flp1, flp2, and flp3 genes was made in H. ducreyi strain 35000HP using Flippase (FLP) recombinase technology as described previously [8,9]. Briefly, two 70 bp primers, P1 and P2, were designed for construction of a cassette (Table 2). The 3' end of each of these primers contained 20 bp complementary to regions 5' and 3' of a spectinomycin cassette flanked by FLP Recognition Target (FRT) sites in pRSM2832 [8]. The 5' portion of the P1 and P2 primers were homologous to regions 5' and 3' of H. ducreyi flp1 and flp3, respectively. PCR of pRSM2832 with P1 and P2 yielded a 2 Kb amplicon that contained the spectinomycin cassette flanked by FRT sites and 50 bp of DNA homologous to regions 5' and 3' of H. ducreyi flp1 and flp3, respectively. This amplicon was electroporated into E. coli DY380 harboring a cosmid size pBeloBAC clone containing the flp operon and flanking DNA. After induction of λ recombinase in this strain, spectinomycin-resistant clones were isolated. One clone was further characterized to demonstrate that the flp1, -2 and -3 genes were replaced with the spectinomycin cassette, with the exception of the flp1 start codon and the terminal 21 bp of the flp3 ORF. The construct was confirmed by sequence analysis.

Table 2. Primers used in this study

The pBeloBAC clone containing the insertion/deletion mutation in the flp genes was used as a template for PCR. The amplicon containing the insertionally inactivated flp1flp2flp3 genes and approximately 500 bp of flanking DNA 5' and 3' to the cassette was ligated into the suicide vector, pRSM2072, and then electroporated into 35000HP. Cointegrates were selected by growth on spectinomycin, then resolved by passage on plates containing spectinomycin and 5-bromo-4-chloro-3-indoly-β-D-galactopyranoside (X-Gal) [24]. Allelic exchange was confirmed by colony PCR.

To make an unmarked mutant, the plasmid, pRSM2975, which contains a temperature sensitive replicon, a kanamycin resistance cassette, and FLP recombinase under the control of the tet repressor, was transformed into the mutant [9]. Transformants were selected and maintained at 32°C on chocolate agar containing kanamycin. The FLP recombinase was induced to catalyze excision of the spectinomycin cassette resulting in a short unmarked ORF in place of the flp1, flp2 and flp3 genes and the plasmid was cured as described previously [9]. Colonies were replica plated on chocolate agar containing no antibiotics, 200 μg/ml of spectinomycin, and 20 μg/ml of kanamycin to identify clones that lost both the temperature sensitive plasmid and the spectinomycin cassette. The plasmid and the spectinomycin cassette were lost in 3/120 (2.5%) of the clones tested. One clone that had a deletion of the expected size by colony PCR was designated 35000HPΔflp1-3.

Lipooligosaccharide (LOS) and outer membrane proteins (OMPs) were prepared from 35000HP and 35000HPΔflp1-3 and were analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis, as described [25]. The growth of parent and mutant in broth cultures were also compared.

RNA isolation and Real Time PCR

Bacterial RNA was prepared from mid-log phase organisms by using TRIzol Reagent (Invitrogen) according to manufacturer's instructions. After isolation, RNA was treated twice with DNaseI (Ambion) for 1 hour at 37°C and then purified by using the RNeasy system (Qiagen). Samples were checked by Agilent analysis. After optimizing primers so that their efficiencies were greater than 95%, we examined the level of transcript expression in RNA isolated from 35000HP and 35000HPΔflp1-3. Using each bacterial RNA, and either the tadA primers (P3 and P4) (Table 2) or the tadG (P5 and P6) primers (Table 2) and SYBR Green, reactions were performed in triplicate using an ABI PRISM 7000 Sequence Detector (Applied Biosystems). Data were expressed as fold change of tadA and tadG in the mutant relative to the parent.

Complementation of 35000HPΔflp1-3

To complement 35000HPΔflp1-3 in trans, the flp1, flp2 and flp3 ORFs were amplified using the P7 primer with a BamH1 linker and the P8 primer with an XhoI linker. The resulting 1.58-kb amplicon was ligated into pCR-XL-TOPO (Invitrogen, Calsbad, Calf.). Transformants were selected on Luria-Bertani plates supplemented with kanamycin (50 μg/ml). The 1.58 kb insert was released from the vector by digestion with BamHI and XhoI, ligated into pLSSK [26], and then transformed into E. coli DH5α. The plasmid was confirmed by restriction mapping and designated pJW1. H. ducreyi 35000HPΔflp1-3 was electroporated with pJW1. As controls, 35000HP and 35000HPΔflp1-3 were electroporated with pLSSK. Transformants were selected on chocolate agar plates containing streptomycin (50 μ/ml) and transformants were saved and designated 35000HPΔflp1-3(pJW1), 35000HP(pLSSK) and 35000HPΔflp1-3(pLSSK).

SDS-PAGE and Western Blot Analysis

Whole cell lysates were prepared from 35000HPΔflp1-3(pJW1), 35000HPΔflp1-3(pLSSK), and 35000HP(pLSSK) and subjected to SDS-PAGE as previously described [27]. In Western Blot analysis, whole cell lysates were probed with rabbit polyclonal sera that bind to Flp1 and Flp2 (kindly provided by Eric J. Hansen) as described elsewhere [4].

Human inoculation protocol

Stocks of 35000HP and 35000HPΔflp1-3 were prepared according to the US Food and Drug Administration guidelines (BB-IND 13046). For the human inoculation protocol, healthy adult male and female volunteers over 21 years of age were recruited for the study. Subjects gave informed consent for participation and for human immunodeficiency virus (HIV) serology, in accordance with the human experimentation guidelines of the U.S. Department of Health and Human Services and the institutional ethics committee of Indiana University-Purdue University of Indianapolis. The experimental protocol, preparation and inoculation of the bacteria, calculation of the estimated delivered dose (EDD), and clinical observations were all done exactly as described previously [10,28]. Subjects were observed until they reached clinical endpoint, which was defined as resolution of all sites, development of a pustule that was either painful or > 6 mm in diameter, or 14 days after inoculation. Subjects were then treated with one dose of oral ciprofloxacin as described [29]. Comparison of papule and pustule formation rates for the two strains were performed using a logistic regression model with generalized estimating equations (GEE) to account for the correlation among sites within the same individual, as previously described [28]. The GEE sandwich estimate for the standard errors was used to calculate 95% confidence intervals (95% CI) for these rates.

To confirm that the strains contained or lacked the flp1flp2flp3 genes, colonies from the inocula, surface cultures and biopsy specimens were replica plated and grown on nitrocellulose filters. Filters were probed with amplicons corresponding to either the fgbA (fibrinogen binder A) gene or the deleted portion of the flp1flp2flp3 genes. The flp1flp2flp3 and the fgbA probes were made using primers P9 and P10 and primers P11 and P12, respectively. Probes were labeled with digoxigenin using the DIG DNA Labeling Kit (Roche Applied Sciences, Penzberg, German) and detected with the DIG Easy Hyb protocol (Roche Applied Sciences) according to the manufacturer's instructions.

Adherence assays

Adherence of bacteria to HFF was measured quantitatively as described previously [4]. Briefly, 24-well tissue culture plates (Costar, Corning, N.Y.) were inoculated with 105 HFF/well and grown to confluence. 35000HPΔflp1-3(pJW1), 35000HPΔflp1-3(pLSSK), and 35000HP(pLSSK) were grown in Columbia broth to an OD660 between 0.4 and 0.6 and harvested by centrifugation. Bacterial pellets were suspended using HFF medium and approximately 106 CFU were added to individual wells containing confluent HFF, centrifuged at 500 × g, and incubated for 2 h at 33°C. After nonadherent bacteria were removed by washing three times with HFF medium, 1 ml of trypsin-EDTA (Invitrogen) solution was added to each well and the plate was incubated for 5 min to liberate the bound bacteria. Serial dilutions of well contents were plated to quantitate HFF-bound bacteria. Percent adherence was calculated as the ratio of HFF-bound bacteria to initial CFU added per well. A paired Student's t test was used for the comparison of strains' abilities to attach to HFF cells.

Microcolony formation assays

Confluent HFF monolayers were prepared and 35000HP(pLSSK), 35000HPΔflp1-3(pLSSK), and 35000HPΔflp1-3(pJW1) were grown as described for the adhesion assays. Bacterial pellets were resuspended in HFF medium to an OD660 of 0.1. Approximately 106 CFU for each strain was added to individual wells of confluent HFF cells, centrifuged at 500 × g for 4 min, and incubated for 24 h at 33°C. Wells were washed three times with 1 ml HFF medium and then stained with crystal violet [0.25% (wt/vol) crystal violet, 20% (vol/vol) methanol, 0.9% (wt/vol) NaCl, 0.02 M Tris-HCl (pH 7.5)] for 20 min at room temperature.

Potential conflicts of Interest

The authors have no conflicts of interest to report.

Authors' contributions

SAC carried out the adherence and microcolony formation assays. JW constructed and characterized the complemented mutant. BB and RSM constructed and characterized the flp1-3 deletion mutant. KRF prepared the bacterial strains used for the human inoculation experiments and participated in the mutant characterization. BWZ and SE prepared the regulatory documents and performed the clinical observations for the human inoculation experiments. BPK performed the statistical analysis. DMJ and RSM participated in the design of the study and drafted the manuscript. All authors read and approved the final manuscript.

Acknowledgements and Funding

We thank Eric Hansen for sharing antibodies used in this work, and S.M. Spinola and M.E. Bauer for their helpful discussions and critical reviews of the manuscript. We thank the volunteers who enrolled in the human challenge study. This work was supported by National Institutes of Health (NIH) National Institute of Allergy and Infectious Diseases (NIAID) grant AI074657 to D.M.J. The human challenge trials were supported by NIH NIAID Public Health Service grant U19 AI31494 and by the Indiana Clinical and Translational Sciences Institute and the Indiana Clinical Research Center (UL RR052761).

References

  1. Schreiner HC, Sinatra K, Kaplan JB, Furgang D, Kachlany SC, Planet PJ, Perez BA, Figurski DH, Fine DH: Tight-adherence genes of Actinobacillus actinomycetemcomitans are required for virulence in a rat model.

    Proc Natl Acad Sci USA 2003, 100:7295-7300. PubMed Abstract | Publisher Full Text | PubMed Central Full Text OpenURL

  2. Tomich M, Planet PJ, Figurski DH: The tad locus: postcards from the widespread colonization island.

    Nat Rev Microbiol 2007, 5:363-375. PubMed Abstract | Publisher Full Text OpenURL

  3. Kachlany SC, Planet PJ, Bhattacharjee MK, Kollia E, DeSalle R, Fine DH, Figurski DH: Nonspecific adherence by Actinobacillus actinomycetemcomitans requires genes widespread in Bacteria and Archaea.

    J Bacteriol 2000, 182:6169-6176. PubMed Abstract | Publisher Full Text | PubMed Central Full Text OpenURL

  4. Nika JR, Latimer JL, Ward CK, Blick RJ, Wagner NJ, Cope LD, Mahairas GG, Munson J, Hansen EJ: Haemophilus ducreyi requires the flp gene cluster for microcolony formation in vitro.

    Infect Immun 2002, 70(6):2965-2975. PubMed Abstract | Publisher Full Text | PubMed Central Full Text OpenURL

  5. Spinola SM, Fortney KR, Katz BP, Latimer JL, Mock JR, Vakevainen M, Hansen EJ: Haemophilus ducreyi requires an intact flp gene cluster for virulence in humans.

    Infect Immun 2003, 71:7178-7182. PubMed Abstract | Publisher Full Text | PubMed Central Full Text OpenURL

  6. Alfa MJ, Stevens MK, Degagne P, Klesney-Tait J, Radolf JD, Hansen EJ: Use of tissue culture and animal models to identify virulence-associated traits of Haemophilus ducreyi.

    Infect Immun 1995, 63:1754-1761. PubMed Abstract | Publisher Full Text | PubMed Central Full Text OpenURL

  7. Baba T, Ara T, Hasegawa M, Takai Y, Okumura Y, Baba M, Datsenko KA, Tomita M, Wanner BL, Mori H: Construction of Escherichia coli K-12 in-frame, single-gene knockout mutants: the Keio collection.

    Mol Syst Biol 2006, 2:2006 0008. PubMed Abstract | Publisher Full Text | PubMed Central Full Text OpenURL

  8. Tracy E, Ye F, Baker BD, Munson RS Jr: Construction of non-polar mutants in Haemophilus influenzae using FLP recombinase technology.

    BMC Molecular Biology 2008, 9:101. PubMed Abstract | BioMed Central Full Text | PubMed Central Full Text OpenURL

  9. Spinola SM, Fortney KR, Baker B, Janowicz DM, Zwickl B, Katz BP, Blick RJ, Munson RS Jr: Activation of the CpxRA system by deletion of cpxA impairs the ability of Haemophilus ducreyi to infect humans.

    Infect Immun 2010, 78:3898-3904. PubMed Abstract | Publisher Full Text | PubMed Central Full Text OpenURL

  10. Janowicz DM, Ofner S, Katz BP, Spinola SM: Experimental infection of human volunteers with Haemophilus ducreyi: 15 years of clinical data and experience.

    J Infect Dis 2009, 199:1671-1679. PubMed Abstract | Publisher Full Text | PubMed Central Full Text OpenURL

  11. Bauer ME, Fortney KR, Harrison A, Janowicz DM, Munson RS Jr, Spinola SM: Identification of Haemophilus ducreyi genes expressed during human infection.

    Microbiology 2008, 154:1152-1160. PubMed Abstract | Publisher Full Text | PubMed Central Full Text OpenURL

  12. Labandeira-Rey M, Brautigam CA, Hansen EJ: Characterization of the CpxRA Regulon in Haemophilus ducreyi.

    Infect Immun 2010, 78:4779-4791. PubMed Abstract | Publisher Full Text | PubMed Central Full Text OpenURL

  13. Labandeira-Rey M, Dodd D, Fortney KR, Zwickl B, Katz BP, Janowicz DM, Spinola SM, Hansen EJ: A Haemophilus ducreyi cpxR deletion mutant is virulent in human volunteers.

    J Infect Dis 2011, 203:1859-1865. PubMed Abstract | Publisher Full Text OpenURL

  14. White CD, Leduc I, Jeter C, Harris C, Elkins C: Haemophilus ducreyi outer membrane determinants, including DsrA, define two clonal populations.

    Infect Immun 2005, 73:2387-2399. PubMed Abstract | Publisher Full Text | PubMed Central Full Text OpenURL

  15. Post DMB, Munson RS Jr, Baker B, Zhong H, Bozue JA, Gibson BW: Identification of genes involved in the expression of atypical lipooligosaccharide structures from a second class of Haemophilus ducreyi.

    Infect Immun 2007, 75:113-121. PubMed Abstract | Publisher Full Text | PubMed Central Full Text OpenURL

  16. Bauer ME, Goheen MP, Townsend CA, Spinola SM: Haemophilus ducreyi associates with phagocytes, collagen, and fibrin and remains extracellular throughout infection of human volunteers.

    Infect Immun 2001, 69:2549-2557. PubMed Abstract | Publisher Full Text | PubMed Central Full Text OpenURL

  17. Bauer ME, Townsend CA, Ronald AR, Spinola SM: Localization of Haemophilus ducreyi in naturally acquired chancroidal ulcers.

    Microbe Infect 2006, 8:2465-2468. Publisher Full Text OpenURL

  18. Fuller TE, Kennedy MJ, Lowery DE: Identification of Pasteurella multocida virulence genes in a septicemic mouse model using signature-tagged mutagenesis.

    Microb Pathog 2000, 29:25-38. PubMed Abstract | Publisher Full Text OpenURL

  19. Harper M, Boyce JD, Wilkie IW, Adler B: Signature-tagged mutagenesis of Pasteurella multocida identifies mutants displaying differential virulence characteristics in mice and chickens.

    Infect Immun 2003, 71:5440-5446. PubMed Abstract | Publisher Full Text | PubMed Central Full Text OpenURL

  20. Kachlany SC, Planet PJ, DeSalle R, Fine DH, Figurski DH, Kaplan JB: flp-1, the first representative of a new pilin gene subfamily, is required for non-specific adherence of Actinobacillus actinomycetemcomitans.

    Mol Microbiol 2001, 40:542-554. PubMed Abstract | Publisher Full Text OpenURL

  21. Labandeira-Rey M, Janowicz DM, Blick RJ, Fortney KR, Zwickl B, Katz BP, Spinola SM, Hansen EJ: Inactivation of the Haemophilus ducreyi luxS gene affects the virulence of this pathogen in human subjects.

    J Infect Dis 2009, 200:409-416. PubMed Abstract | Publisher Full Text | PubMed Central Full Text OpenURL

  22. Al-Tawfiq JA, Thornton AC, Katz BP, Fortney KR, Todd KD, Hood AF, Spinola SM: Standardization of the experimental model of Haemophilus ducreyi infection in human subjects.

    J Infect Dis 1998, 178:1684-1687. PubMed Abstract | Publisher Full Text OpenURL

  23. Janowicz DM, Fortney KR, Katz BP, Latimer JL, Deng K, Hansen EJ, Spinola SM: Expression of the LspA1 and LspA2 proteins by Haemophilus ducreyi is required for virulence in human volunteers.

    Infect Immun 2004, 72:4528-4533. PubMed Abstract | Publisher Full Text | PubMed Central Full Text OpenURL

  24. Fulcher RA, Cole LE, Janowicz DM, Toffer KL, Fortney KR, Katz BP, Orndorff PE, Spinola SM, Kawula TH: Expression of Haemophilus ducreyi collagen binding outer membrane protein NcaA is required for virulence in swine and human challenge models of chancroid.

    Infect Immun 2006, 74:2651-2658. PubMed Abstract | Publisher Full Text | PubMed Central Full Text OpenURL

  25. Fortney KR, Young RS, Bauer ME, Katz BP, Hood AF, Munson RS Jr, Spinola SM: Expression of peptidoglycan-associated lipoprotein is required for virulence in the human model of Haemophilus ducreyi infection.

    Infect Immun 2000, 68:6441-6448. PubMed Abstract | Publisher Full Text | PubMed Central Full Text OpenURL

  26. Wood GE, Dutro SM, Totten PA: Target cell range of Haemophilus ducreyi hemolysin and its involvement in invasion of human epithelial cells.

    Infect Immun 1999, 67:3740-3749. PubMed Abstract | Publisher Full Text | PubMed Central Full Text OpenURL

  27. Spinola SM, Griffiths GE, Bogdan JA, Menegus MA: Characterization of an 18,000 molecular-weight outer membrane protein of Haemophilus ducreyi that contains a conserved surface-exposed epitope.

    Infect Immun 1992, 60:385-391. PubMed Abstract | Publisher Full Text | PubMed Central Full Text OpenURL

  28. Spinola SM, Bong CTH, Faber AL, Fortney KR, Bennett SL, Townsend CA, Zwickl BE, Billings SD, Humphreys TL, Bauer ME, et al.: Differences in host susceptibility to disease progression in the human challenge model of Haemophilus ducreyi infection.

    Infect Immun 2003, 71:6658-6663. PubMed Abstract | Publisher Full Text | PubMed Central Full Text OpenURL

  29. Banks KE, Fortney KR, Baker B, Billings SD, Katz BP, Munson RS Jr, Spinola SM: The enterobacterial common antigen-like gene cluster of Haemophilus ducreyi contributes to virulence in humans.

    J Infect Dis 2008, 197:1531-1536. PubMed Abstract | Publisher Full Text OpenURL