|Home | About | Journals | Submit | Contact Us | Français|
Haemophilus ducreyi 35000HP contains a homolog of the luxS gene, which encodes an enzyme that synthesizes the AI-2 autoinducer in other gram-negative bacteria. H. ducreyi 35000HP produced AI-2 that functioned in a Vibrio harveyi-based reporter system. An H. ducreyi luxS mutant was constructed by insertional inactivation of the luxS gene and lost the ability to produce AI-2. Provision of the H. ducreyi luxS gene in trans partially restored AI-2 production by the mutant. The luxS mutant was compared to its parent for virulence in the human challenge model of experimental chancroid. The pustule formation rate in five volunteers was 93.3% (95% CI, 81.7, 99.9) at 15 parent sites and 60.0% (95% CI, 48.3, 71.7) at 15 mutant sites (one-tailed P = 0.0002). Thus, the luxS mutant was partially attenuated for virulence. This is the first report of AI-2 production contributing to the pathogenesis of a genital ulcer disease.
Quorum sensing refers to the process whereby bacteria control their own gene expression based on the local density of their population . Over the past twenty years, quorum sensing has been shown to regulate virulence factors and influence biofilm formation in several important bacterial pathogens . This regulatory system involves the production of a signaling molecule, called an autoinducer, whose extracellular concentration increases as the cell density increases. When a concentration threshold is attained, the bacterial population alters its gene expression in unison . Several different cell density-dependent regulatory systems have been described, and three of these have been well characterized in gram-negative bacteria. The LuxR-LuxI system involves a N-acyl-homoserine lactone autoinducer. The second system involves autoinducer-2 (AI-2), first identified as a furanosyl-borate diester in Vibrio harveyi . The third system uses AI-3, a compound chemically distinct from AI-2 . The first system is involved mostly in intraspecies communication whereas the second and third systems are involved in interspecies communication . Inactivation of luxS, a key component in the AI-2-based system, has been shown to affect the expression of virulence factors in several bacteria including V. cholerae , Aggregatibacter (formerly Actinobacillus) actinomycetemcomitans , and enterohemorrhagic Escherichia coli . Recently, LuxS has been shown to be functional in several members of the Pasteurellaceae, including Actinobacillus pleuropneumoniae [6, 8], A. actinomycetemcomitans  and Mannheimia haemolytica [9, 10].
Haemophilus ducreyi, a member of the Pasteurellaceae and the etiologic agent of chancroid, carries in its genome (GenBank accession number NC002940) a homologue that is 72% identical to the V. harveyi luxS gene. Although production of AI-2 by H. ducreyi has not been demonstrated, the presence of luxS and the facts that H. ducreyi forms microcolonies , a key feature of biofilm formation  in vitro, and aggregates during human infection , suggested the possible existence of a quorum sensing mechanism for gene regulation in H. ducreyi.
To study the pathogenesis of H. ducreyi infection, we developed an experimental model of infection in human volunteers [14, 15]. The model involves the injection of the human-passaged H. ducreyi strain 35000HP and its derivatives into the skin of the upper arm [16, 17]. There are gender and host effects on outcome in the model . In mutant-parent comparison trials, volunteers are inoculated with multiple doses of 35000HP and mutants made in the 35000HP background on opposite arms and serve as their own controls for gender and host effects . Of twenty mutants tested to date, eleven are virulent (form pustules at doses similar to the parent) [17, 19, 20], three are partially attenuated (form pustules at doses two to three times that of the parent or at lower rates at doses equivalent to the parent) [21-23], and six are attenuated (unable to form pustules even at doses 10-fold that of the parent) [24-29]. Thus, the model discriminates among H. ducreyi gene products for their abilities to contribute to the development of pustules.
In the present study, we report that H. ducreyi synthesizes an AI-2-like molecule. We show that inactivation of the H. ducreyi luxS gene decreased the virulence of this pathogen in human volunteers.
H. ducreyi strains (table 1) were routinely cultured on chocolate agar (CA) plates, and incubated at 33°C in a humidified atmosphere containing 95% air and 5% CO2. For in vitro experiments, strains were grown at 33°C in a gyratory water bath at 100 rpm in a Columbia broth-based (CB) medium as described . H. ducreyi cells used as inocula in the human challenge model were grown as described . Plasmids used in this study are listed in table 1. When appropriate, media was supplemented with chloramphenicol or kanamycin at final concentrations of 1 μg/ml and 30 μg/ml, respectively. V. harveyi strains were grown in AB medium . E. coli DH5α and TOP10′ strains were used for general cloning manipulations and were grown in Luria-Bertani medium supplemented with kanamycin (50 μg/ml) or chloramphenicol (30 μg/ml) when appropriate for maintenance of plasmids. Plasmid constructs were transformed into and isolated from E. coli HB101 before they were electroporated into H. ducreyi.
The luxS gene was amplified by PCR from H. ducreyi 35000HP chromosomal DNA using oligonucleotide primers luxS-7 (5′-ATGCGGAGCAGTTGACTAAC-3′) and luxS-8 (5′-ATTCTGCACTACGAATACCC-3′). The resulting 3.6-kb amplicon was ligated into pCRII Blunt TOPO to obtain pRB302. The ΔEcat cartridge  was ligated into a unique SwaI site in luxS in pRB302 to obtain pRB303. E. coli HB101 was transformed with pRB303; the plasmid was purified and linearized with SmaI. The linearized construct was gel-purified and electroporated into 35000HP. Transformants were allowed to recover in CB for 5 hours, after which they were transferred to CA plates containing chloramphenicol (1 μg/ml). Following an incubation period of 72-144 hours, colonies were screened for the presence of a mutated luxS gene by direct single colony PCR amplification using oligonucleotide primers luxS-13 (5′-CTTTTCTCTGGTGAGATCATTACAG-3′) and luxS-14 (5′-TTTAACGCCATCGACTAGAGTC-3′). The luxS mutant was designated 35000HP.303.
To complement 35000HP.303, a fragment that included the luxS gene together with 657-nt upstream and 164-nt downstream DNA was used. Because luxS appears to be in an operon with at least two other genes, to ensure transcription, the promoter region (-1 nt to -150 nt) of the pACYC184 cat gene  was inserted in front of the luxS gene (at -1 nt) using overlapping extension PCR . Three amplicons were used in this procedure. Fragment 1 was obtained by amplification of the luxS upstream region using primers luxS-BamHI-F (5′-ACGCGGATCCATGCTAAAAAACAGATAAAATCAGATAAAA-3′) and luxS-2 (5′-GTTGCCCCAGGGCTTCCCGGTATCAACATTTTACTCCAAATTATAT-3′). Fragment 2 involved amplification of the cat promoter region using primers catP-3 (5′-GTTGATACCGGGAAGCCCTGGGCCAAC-3′) and catP-4 (5′-TTTAGCTTCCTTAGCTCCTGAAAA-3′). Fragment 3 required amplification of luxS and its downstream region using primers luxS-5 (5′-TTTTCAGGAGCTAAGGAAGCTAAAATGCCTTTATTAGATAGTTTT-3′) and luxS-BamHI-R (5′-ACGCGGATCCCTTTTCTCTGGTGAGATCATTACAG-3′). The underlined sequences indicate BamHI sites and sequences in bold indicate regions complementary to fragment 2. All three fragments were gel-purified and equivalent amounts were used together as templates in the final overlapping extension PCR using primers luxS-BamHI-F and luxS-BamHI-R. The resulting amplicon was ligated as a BamHI fragment into plasmid pACYC177 , and the resultant construct, designated pML132, was transformed into 35000HP.303 as described above. Nucleotide sequence analysis verified the absence of PCR-induced mutations in the final cloned amplicon. The complemented mutant was designated 35000HP.303(pML132).
Chromosomal DNA from 35000HP and 35000HP.303 were digested with EcoRV and used in Southern blot analysis. The probe for the luxS gene was obtained by PCR, using the primers luxS-9 (5′-TCGATTTAACCATTCTTCATC-3′) and luxS-10 (5′-TATTGATATTTCACCGATGG-3′) to amplify a 283-bp fragment from 35000HP chromosomal DNA. The probe for the cat cartridge was obtained by using the primers DeltaEcat(F) (5′-CCGTTTTTATCAGGCTCTGG-3′) and DeltaEcat(R) (5′-TTTTGCCGTTACGCACCAC-3′) to amplify a ~1.3-kb fragment from pACYC184ΔEcat containing the cat cartridge. Both probes were labeled by using the PCR DIG Labeling Mix (Roche Applied Sciences, Indianapolis, IN).
A 500 μl volume of stop buffer [200 mM Tris-Cl (pH 8), 20mM EDTA, and 20mM sodium azide] was added to 10 ml of mid-logarithmic phase bacterial culture immediately prior to collection by centrifugation. Total RNA was isolated from 35000HP immediately upon harvest by using a RiboPure-Bacteria Kit (Ambion, Austin, TX). After isolation, RNA was treated with DNaseI (Ambion) for 1 h at 37°C and purified using the RNeasy system (Qiagen). Primers used for RT-PCR (figure 1A) were designed using the 35000HP genome sequence and had the following sequences: HD0369-HD0370 (F) (5′-GTTATACGATGGATATTTCT-3′); HD0369-HD0370 (R) (5′-CTCCATATTAGGGCGACAAA-3′); HD0370-HD0371 (F) (5′-ACACAAATTCCTGAGCTGAA-3′) and HD0370-HD0371 (R) (5′-ATATAAAAGGATAAATACGC-3′). Briefly, 2 μg RNA was reverse-transcribed using TaqMan® Reverse Transcription Reagents (Applied Biosystems) according to the manufacturer’s instructions. The resulting cDNA was used as PCR template with the primer sets above. A reaction lacking RT was used as a PCR template to control for contaminating DNA. Real-time RT-PCR  was used to measure relative expression levels of HD0371 in 35000HP and 35000HP.303, using oligonucleotide primers 5′-ATCGAGATTGGACGTTGGTA-3′ and 5′-CTCGTCTAGCGCTTTCACC-3′.
Preparation of cell-free culture supernatants was performed as previously described . Briefly, H. ducreyi strains were grown in CB overnight and used to inoculate CB to an initial OD600=0.05. Portions (500 μl) of the culture were harvested every hour for 16 h. Supernatants were prepared by centrifugation at 2,500 × g for 5 min. These fluids were then sterilized with a 0.2 μm pore size filter and stored at -20°C until used. As a positive control, cell-free culture supernatant was also prepared from V. harveyi strain BB120 as described . The autoinducer bioassay was performed as follows: 10 μl portions of cell-free culture supernatants were added to a 96-well sterile black/clear bottom microtiter plate. The V. harveyi reporter strain BB170 was grown for 14-16 h in AB medium to a density of 2.4 × 109 CFU/ml and diluted 1:5000 in fresh AB medium. A 90-μl portion of this cell suspension was added to each of the wells containing the supernatants. Negative control wells contained 90 μl of cells and 10 μl of either CB or AB medium. The microtiter plate was shaken at 175 rpm at 30°C and light production was measured after 5 h in a TECAN SPECTRAFluor Plus fluorometer set to luminescence mode with a 200 ms integration time and a gain setting of 100.
Healthy 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 Review Board 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 done exactly as described previously [14, 15]. Stocks of 35000HP and 35000HP.303 used for the human inoculation experiments were prepared according to FDA guidelines under IND#BB-IND 13064. 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 > 4 mm in diameter, or 14 days after inoculation. At endpoint, subjects were treated with one dose of oral ciprofloxacin.
The trial was designed to test whether the luxS mutant was impaired in its ability to cause pustules relative to the parent in a multi-stage dose-ranging study with a minimum of 2 stages . An EDD of 90 CFU of 35000HP causes a papule formation rate of 95% and a pustule formation rate of 50% in naive subjects. We attempted to inoculate each subject in the first group with 90 CFU of 35000HP at 3 sites on one arm and with 45, 90 and 180 CFU of 35000HP.303 on the other arm so that the dose of the mutant is in a range that causes pustules if the mutant is virulent. If inoculation of the mutant and parent causes pustules in the first group of subjects, the second group is inoculated with similar doses. 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 . The GEE sandwich estimate for the standard errors was used to calculate 95% confidence intervals (95% CI) for these rates.
To confirm that the inocula contained or lacked the cat cartridge and that the cat cartridge was not lost by the mutant during the course of infection, individual colonies from the inocula, surface cultures, and biopsy specimens were scored for susceptibility to chloramphenicol on chloramphenicol-containing CA plates as described .
Analysis of the H. ducreyi genome identified a 509-nt ORF encoding a LuxS homolog. Although the presence of luxS suggested that H. ducreyi is capable of producing AI-2, homologues of the known V. harveyi AI-2 receptor and signal transduction genes (luxP, luxQ, luxU, and luxO) have not been identified in H. ducreyi. The H. ducreyi genes HD0369 and HD0371, which immediately flank luxS (figure 1A), encode a paraquat-inducible protein B and a predicted membrane-associated, metal-dependent hydrolase, respectively. RT-PCR analysis showed that all three genes (HD0369-luxS-HD0371) are transcriptionally linked in 35000HP (figure 1A).
To determine whether the H. ducreyi luxS homolog might be involved in virulence expression, we constructed a luxS mutant by inserting a cat cartridge into the luxS ORF as described in Materials and Methods. PCR analysis of the luxS gene in this mutant indicated that it was approximately 1.3 kb larger than the wild-type luxS gene (figure 1B), consistent with the insertion of the 1.3 kb cat cartridge. The insertion of a single antibiotic cartridge in the luxS ORF was confirmed by Southern blot analysis. When EcoRV-digested chromosomal DNA from the luxS mutant was probed with a 283-bp fragment of luxS, a 6.8 kb fragment hybridized with this probe (figure 1C). This fragment was 1.3 kb larger than the 5.5 kb fragment of EcoRV-digested 35000HP chromosomal DNA that hybridized with this same probe (figure 1C). A 6.8 kb fragment of EcoRV-digested 35000HP.303 chromosomal DNA also hybridized with a probe for ΔEcat (figure 1C) whereas chromosomal DNA from 35000HP failed to hybridize with the ΔEcat probe (figure 1C). Real-time RT-PCR analysis indicated that expression of the downstream ORF HD0371 was only slightly reduced (log10 of relative expression level = -0.10 compared to 35000HP) in 35000HP.303; expression of the encoded protein was not measured. 35000HP and 35000HP.303 had similar outer membrane protein and lipooligosaccharide expression patterns as analyzed by SDS-PAGE and had similar growth rates in the broth used to prepare inocula for human challenge experiments (data not shown).
To investigate if H. ducreyi-conditioned medium had AI-2 activity, a bioluminescence assay was performed using the V. harveyi reporter strain BB170. V. harveyi BB170 cannot synthesize its own AI-2; only AI-2 present in the conditioned medium should induce luminescence . Cell-free culture supernatants collected during the growth of 35000HP and 35000HP.303 were incubated with BB170. V. harveyi BB120-conditioned medium was used as a positive control. 35000HP produced significant amounts of AI-2 activity when compared to 35000HP.303 (figure 2). Peak activity was observed at mid-exponential phase and then decreased during stationary phase, which correlates well with results obtained with other bacteria that produce AI-2 . Expression of luxS in trans partially restored the ability of 35000HP.303 to induce luminescence in this assay (figure 2).
Eight healthy adults (four males, four females; 6 Caucasian, 2 Black; age range 25 to 45; mean age ± standard deviation (SD), 38 ± 9 years) volunteered for the study. Two subjects were excluded because they did not meet inclusion criteria. One subject withdrew on the day of inoculation. Three subjects (volunteers 318, 319 and 322) were inoculated in the first iteration and two subjects (volunteers 323 and 328) were inoculated in the second iteration. In the first iteration, each subject was infected with 143 CFU of 35000HP at three sites on one arm and with 59, 118 and 235 CFU of 35000HP.303 on the other arm (table 2). Pustules formed at 8 of 9 parent sites and 6 of 9 mutant sites. Because the mutant and the parent both formed pustules, we repeated the experiment on a second group of subjects. In the second iteration, 2 volunteers were inoculated with 101 CFU of 35000HP on one arm and 93, 186 and 372 CFU of 35000HP.303 on the other arm. Pustules formed at 6 of 6 parent sites and 3 of 6 mutant sites (table 2).
Overall, the mean EDD ± SD for the parent was 126 ± 21 CFU and for the mutant was 169 ± 104 CFU. By 24 h after inoculation, the mean area ± SD of papules at parent sites was 31.0 ± 10.7 mm2 and at mutant sites was 23.9 ± 12.2 mm2 (P= 0.11). The pustule formation rate was 93.3% (95% CI, 81.7, 99.9) at parent sites and 60.0% (95% CI, 48.3, 71.7) at mutant sites (one-tailed P = 0.0002).
Of the 15 sites that were inoculated with the parent, 12 (80.0%) yielded at least one positive surface culture, while 7 of 15 mutant sites (46.7%) yielded a positive surface culture. Of five parent sites and 5 mutant sites that were biopsied and cultured, 4 yielded H. ducreyi (table 3); the yields were not statistically different. Histological examination was done on all biopsy specimens as described [19, 20]. All biopsies contained micropustules composed of neutrophils within the epidermis and a dermal infiltrate of CD4 positive perivascular mononuclear cells. No major differences were observed between mutant and parent pustules (data not shown).
For the parent and mutant broth cultures used to prepare the inocula, all 71 parent colonies and all 72 mutant colonies tested had the expected antibiotic susceptibility (parent, chloramphenicol-sensitive; mutant, chloramphenicol-resistant). All colonies obtained from surface cultures and biopsy specimens from parent sites (n = 387 and n = 156, respectively) and mutant sites (n = 319 and n = 114, respectively), had the expected antibiotic susceptibility.
Production of AI-2 by luxS has been established for many bacterial species . While the involvement of the luxS/AI-2 system in quorum-sensing remains to be proven for many of them, inactivation of luxS results in altered virulence phenotypes in some pathogens [5-7]. In the present study, we have shown (1) that H. ducreyi contains a luxS homolog, (2) that this bacterium is capable of producing an AI-2-like molecule, and (3) that the inactivation of luxS results in modestly reduced infectivity in the human challenge model.
When tested in the V. harveyi-based bioluminescence assay, cell-free culture supernatants from H. ducreyi 35000HP induced luminescence while 35000HP.303 had considerably less AI-2 activity, indicating that H. ducreyi produces an AI-2-like molecule that is dependent on the presence of LuxS. Other Pasteurellaceae family members have a functional LuxS/AI-2 regulatory system [8-10, 37]. When luxS is inactivated in these bacteria, the virulence of the organisms is altered. An H. influenzae luxS mutant has increased virulence in a chinchilla model of middle ear infection  while A. pleuropneumoniae and M. haemolytica A1 mutants have reduced virulence in mouse  and calf infection  models, respectively. In the present study, we showed that inactivation of the luxS gene resulted in a reduced pustule formation rate when compared to the parental strain, indicating that this mutant has a partially attenuated phenotype. This is the fourth report of a partially attenuated mutant in 35000HP.
The human model closely simulates the first two weeks of natural chancroid. Lesions are not allowed to progress beyond the pustular stage to ulcers, which are associated with lymphadenitis. In pustules, H. ducreyi forms aggregates and co-localizes with collagen, fibrin, neutrophils and macrophages and is extracellular [13, 38]. It is possible that H. ducreyi uses the production of AI-2 to alter its gene regulation profile and to aid in the initiation of infection. Since H. ducreyi maintains similar relationships with host cells in experimental infection and naturally acquired chancroidal ulcers , the expression of luxS may also contribute to the ulcerative stage of disease.
We insertionally inactivated luxS gene by insertion of a cat cartridge. Complementation of the luxS mutant in trans partially restored AI-2 activity. The lower AI-2 activity obtained with the complemented luxS mutant (figure 2) may be the result of the reduced extent of growth caused by the presence of kanamycin in the growth medium. We are precluded by biosafety considerations from testing complemented mutants in human volunteers. According to the human challenge protocols, we are only required to show, by Southern blot and PCR analyses, that the expected allelic exchange had occurred in the mutant. Subsequent to performing the human challenge trials, we used PCR to amplify the mutated luxS gene and some flanking DNA from the chromosome of the luxS mutant. Nucleotide sequence analysis showed two changes in the flanking ORFs, resulting in one amino acid substitution (S762N) in the HD0369 protein and one amino acid substitution (I28T) in the HD0371 protein.
We examined the role of LuxS in the virulence of 35000HP, a class I strain. There are two classes of H. ducreyi strains, which express different variants of several outer membrane proteins and oligosaccharide components of LOS . Whether LuxS plays a role in the virulence of class II strains is unknown.
To our knowledge, this is the first report of a genital ulcer disease pathogen producing an AI-2-like molecule that likely plays a role in virulence. Further studies to determine (1) whether this autoinducer molecule is involved in gene regulation in H. ducreyi, (2) the mechanism by which the luxS/AI-2 system contributes to virulence of H. ducreyi 35000HP, and (3) whether the luxS gene product plays a role in virulence of class II H. ducreyi strains  are in progress.
We thank Bonnie Bassler for providing the V. harveyi strains and Bruce Green of Wyeth Vaccines for providing the ΔEcat cartridge. We thank Shelia Ellinger for preparation of regulatory documents and the volunteers who participated in the trial.
Financial support: U.S. Public Health Service grants AI31494 to S.M.S, AI32011 to E.J.H. and AI07637 to D.M.J. NIH (grant UL RR025761 to the Indiana Clinical Research Center, a component of the Indiana Clinical and Translational Sciences Institute) and MO1RR00750 to the GCRC at Indiana University to support the human challenge trials.
Potential conflicts-of-interest: None