Haemophilus ducreyi Is Susceptible to Protegrin

Abstract

Protegrins, potent antimicrobial peptides found in porcine leukocytes, have activity against the sexually transmitted pathogens Neisseria gonorrhoeae, Chlamydia trachomatis, and human immunodeficiency virus type 1. We tested synthetic protegrin 1 (PG-1) for activity against nine isolates of Haemophilus ducreyi, the etiologic agent of chancroid. The test organisms included CIP 542 (the type strain), 35000HP (a human-passaged variant of 35000), 35000HP-RSM2 (an isogenic d-glycero-d-manno-heptosyltransferase mutant of 35000HP), and six clinical isolates. The isolates were epidemiologically unrelated, represented three HindIII ribotypes, and had varying antimicrobial resistance patterns. In bactericidal assays, five isolates were rapidly killed by synthetic PG-1. In radial diffusion assays, all nine isolates were exquisitely sensitive to PG-1. These data highlight the potential of protegrins for development as topical agents to prevent many sexually transmitted diseases, including chancroid.

Haemophilus ducreyi causes chancroid, a genital ulcer disease common in developing countries (10, 20). In a process called epidemiologic synergy, H. ducreyi and the human immunodeficiency virus (HIV) facilitate the transmission of each other (1, 16, 18, 21). The impact of chancroid and other sexually transmitted diseases (STDs) on heterosexually acquired HIV infection has increased interest in the development of topical microbicides that are active against all the major STD pathogens.

Protegrins are cysteine-rich antimicrobial peptides derived from porcine leukocytes (7, 24). Protegrins contain 16 to 18 amino acid residues and a β-sheet structure that is stabilized by two intramolecular cysteine disulfide bonds. The maintenance of the β-sheet structure by the disulfide bonds is required for the retention of antimicrobial activity in solutions containing salt concentrations similar to those found in extracellular fluid (4). Protegrins inactivate Neisseria gonorrhoeae, Chlamydia trachomatis, HIV type 1 (HIV-1) (11, 17, 22), and herpes simplex virus type 2 (unpublished data). Protegrins kill gram-negative bacteria by binding to lipids and permeabilizing the outer and inner membranes. In studies performed with derivatives of protegrins, a 12-mer containing one disulfide bond retains almost all activity against N. gonorrhoeae (12). However, optimal activity against C. trachomatis requires both disulfide bonds (23).

The activity of protegrins against N. gonorrhoeae, C. trachomatis, herpes simplex virus type 2, and HIV-1 makes them excellent candidates for use as topical agents against STDs. Here we tested the ability of synthetic protegrin 1 (PG-1) to kill several H. ducreyi isolates, including the type strain, an isolate recovered from an experimentally infected human subject and its isogenic lipooligosaccharide (LOS) mutant, and six recent clinical isolates.

(This work was presented at the International Congress of Sexually Transmitted Disease in Seville, Spain, October 1997.)

MATERIALS AND METHODS

Bacteria.

The origins and years of isolation of the H. ducreyi strains are listed in Table 1. H. ducreyi CIP 542 is the type strain (5). H. ducreyi 35000HP (human passaged) was recovered from a subject who was experimentally infected with 35000 (ATCC 33922) for 13 days (15). H. ducreyi 35000HP-RSM2 (kindly provided by Robert S. Munson, Jr.) is an isogenic mutant of 35000HP and contains an Ωkan insertion in losB, which encodes d-glycero-d-manno-heptosyltransferase (3). H. ducreyi 35000HP-RSM2 produces a truncated LOS that terminates in a single glucose attached to a heptose trisaccharide core and 2-keto-3-deoxyoctulosonic acid (KDO) and is identical in structure to a Tn916 derivative of 35000, designated 1381 (3). H. ducreyi 35000HP-RSM2 and 35000HP have identical outer membrane protein profiles and growth rates, and both isolates are resistant to normal human serum (unpublished observations). The HMC series are recent clinical isolates of H. ducreyi with diverse geographic origins. Except as indicated below, the isolates were routinely grown on chocolate agar supplemented with 1% IsoVitaleX at 35°C in a 5% CO₂ atmosphere. For ribotyping and determination of MICs, the chocolate agar was supplemented with 5% fetal bovine serum (FBS) as described previously (6).

TABLE 1.

H. ducreyi strains used in this study

Strain	Origin or source	Yr isolated	HindIII ribotype	MIC (μg/ml)a								PG-1 x-interceptb
				PC	TC	KM	CM	EM	AZ	CIP	CFT
35000HP	Spinola lab	1996	1	0.5	1	4	0.5	0.06	0.03	0.015	0.004	1.9 ± 0.7
35000HP-RSM2	Munson lab	1997	NDc	ND	ND	>16	ND	ND	ND	ND	ND	1.0 ± 0.6
CIP 542	Hanoi	1954	1	0.5	1	4	0.5	0.06	0.015	0.03	0.004	1.9 ± 0.5
HMC 46	Kenya	1995	4	>128	32	8	16	0.01	0.004	0.008	0.002	0.7 ± 0.4
HMC 48	Bahamas	1995	2	>128	32	4	16	0.03	0.015	0.015	0.001	2.3 ± 0.2
HMC 49	Mississippi	1995	1	1	1	8	0.5	0.06	0.015	0.015	0.004	2.3 ± 0.6
HMC 53	Mississippi	1995	2	>128	32	4	16	0.06	0.008	0.03	0.002	5.0 ± 1.3
HMC 56	Dominican republic	1995	4	>128	32	2	1	0.02	0.008	0.03	0.001	0.6 ± 0.6
HMC 88	Seattle	1995	2	>128	64	>16	16	0.03	0.008	0.015	0.002	3.6 ± 1.5

^aPC, penicillin; TC, tetracycline; KM, kanamycin; CM, chloramphenicol; EM, erythromycin; AZ, azithromycin; CIP, ciprofloxacin; CFT, ceftriaxone. 

^bValues are means ± standard deviations, in micrograms per milliliter. 

^cND, not determined. 

Ribotyping.

The H. ducreyi isolates were ribotyped by a modification of the method of Sarafian and coworkers (13), except that rRNA probes were prepared by using a reverse transcriptase kit according to the manufacturer’s directions (Gibco-BRL, Gaithersburg, Md.) rather than end labeling. The bands of the isolates were compared with those of the 13 known HindIII ribotypes (13) (unpublished data).

MICs.

The H. ducreyi isolates were tested for their susceptibility to the following antibiotics: erythromycin (Sigma Chemical Co., St. Louis, Mo.), azithromycin (Pfizer, Inc., Groton, Conn.), ceftriaxone (Hoffmann-La Roche, Inc., Nutley, N.J.), and ciprofloxacin (Bayer, Westhaven, Conn.), penicillin (Eli Lilly, Indianapolis, Ind.), tetracycline (Sigma), chloramphenicol (Sigma), and kanamycin (Sigma). MICs were determined by a plate dilution method described previously for H. ducreyi (6) with modifications in the inoculum preparation to reduce clumping of the bacteria. Bacteria were grown overnight on freshly prepared chocolate agar-FBS plates (GC agar base, 1% hemoglobin, 1% IsoVitaleX, and 5% FBS) and adjusted to the density of a 0.5 McFarland standard. Approximately 10⁴ CFU were applied to chocolate agar-FBS plates containing serial twofold dilutions of antibiotics by using a Steers replicator. All plates were incubated at 35°C in an atmosphere of 3% CO₂ and analyzed after 24 and 48 h of incubation; there were no differences in the results at these two time points.

Peptides.

Synthetic PG-1 was C-terminally amidated, prepared with Fmoc (9-fluorenylmethoxycarbonyl) chemistry (SynPep, Dublin, Calif.) and purified in our laboratory (12). Synthetic PG-1 was identical to native PG-1 in mass, net charge, behavior in sodium dodecyl sulfate-acid urea polyacrylamide gel electrophoresis, and activity against a panel of gram-positive and gram-negative bacteria (data not shown).

Bactericidal assays.

H. ducreyi was grown overnight at 34°C in brain heart infusion (BHI) (Difco Laboratories, Detroit, Mich.) broth containing 0.1% soluble starch (Fisher Scientific, Itasca, Ill.), 1% IsoVitaleX, and 50 μg of hemin (Aldrich Chemical Co., Milwaukee, Wis.) per ml. The cells were diluted in fresh broth and grown to mid-log phase. The cells were harvested, rinsed three times, and suspended in phosphate-buffered saline (PBS) containing 1% BHI (PBS-BHI).

PG-1 was dissolved in 0.01% acetic acid. In each assay, 10⁶ to 10⁷ CFU of bacteria per ml in 9 volumes of PBS-BHI were incubated with 1 volume of PG-1 or 1 volume of 0.01% acetic acid lacking PG-1. Time course studies were done with a protegrin concentration of 50 μg/ml in a final volume of 200 μl, and the number of CFU was determined after 15 min, 1 h, and 2 h of incubation. Dose response studies were done with protegrin concentrations ranging from 1.6 to 50 μg/ml in a final volume of 100 μl, and the number of CFU was determined after 1 h of incubation. All assays were done at 34°C. The number of CFU was determined by plating in triplicate 10 μl of undiluted sample and 10-fold serial dilutions of the sample. Bacterial killing was expressed as the log₁₀ of the CFU per milliliter in the acetic acid control wells divided by the log₁₀ of the CFU per milliliter in the protegrin-treated wells for each data point. In these assays, the lowest detectable concentration of bacteria was 10² CFU/ml. Thus, maximal detectable killing was expressed as ≥4 log₁₀ CFU. All assays were performed three to five times.

To examine whether antibiotic carryover occurred in the bactericidal assays, 10³ CFU of 35000HP were added to wells containing either 50 μg of protegrin per ml or the acetic acid control in a final volume of 100 μl, and 10 μl was immediately plated. There were no differences in the CFU recovered between the protegrin-treated and control wells.

Bacterial viability staining.

Staining for bacterial viability was done with the LIVE/DEAD BacLight Bacterial Viability Kit (Molecular Probes, Inc., Eugene, Oreg.). H. ducreyi 35000HP was incubated with protegrin (50 μg/ml) or the acetic acid control as described above for 15 min. The samples (200 μl) were diluted in 1 ml of PBS, centrifuged, suspended in 100 μl of PBS, and stained for 15 min at room temperature in the dark with 0.3 μl of 3.34 mM STYO 9 green and with 1 μl of 2.0 mM propidium iodide. The samples were viewed with a fluorescein-rhodamine dual filter and photographed.

Radial diffusion assays.

Radial diffusion assays were done by a modification of methods described previously (8, 11, 16). Underlay gels contained 1.5 g of proteose peptone (Difco), 4 g of K₂HPO₄, 1 g of KH₂PO₄, 5 g of NaCl (all salts purchased from Fisher Scientific), 1 g of soluble starch, and 10 g of A-6013 agarose (Sigma Chemical Co.) per liter. Agarose was included in the nutrient-poor underlay gel to avoid electrostatic interactions between the protegrin and the polyionic components of agar (16). Nutrient-rich overlay gels were made from the same components as the underlay gels, except that the concentration of proteose peptone was 15 g/liter, and 10 g of agar (Sigma) was substituted for the agarose. The overlay gels also contained 2% IsoVitaleX, 20% FBS (HyClone Laboratories, Logan, Utah), and 400 mg of catalase (Sigma) per liter. Catalase was used as a porphyrin source so that the plates were clear (19) and had easily readable zones of inhibition.

Ten milliliters of the underlay gel was mixed with approximately 10⁶ to 10⁷ CFU of H. ducreyi that had been grown on chocolate agar plates overnight and poured into a 100- by 15-mm petri dish. Seven evenly spaced wells were made in the gel with a 3-mm punch forceps, and the agarose plugs were removed with a pipette tip. PG-1 was dissolved and serially diluted in 0.01% acetic acid and 0.1% bovine serum albumin (Reheis Chemical Co., Phoenix, Ariz.), and 8 μl of each dilution was added to each well. The plates were incubated at 35°C in a 5% CO₂ atmosphere for 3 h, and 10 ml of the overlay gel was added to the plates. The plates were incubated for an additional 48 to 72 h, and the clear zones surrounding the wells were measured. The diameter of clearing was expressed in units (0.1 mm = 1 U) and was calculated after the diameter (3 mm = 30 U) of the well had been subtracted as described previously (8, 16). In each experiment, an individual isolate was tested on duplicate plates. Each experiment was repeated two to four times.

The susceptibilities of 35000HP and 35000HP-RSM2 were determined four times concurrently. The results were compared by using a mixed-effects analysis of variance model to adjust for differences in experiments performed on different days.

RESULTS

Classification and antimicrobial susceptibility of the H. ducreyi isolates.

The origins and years of isolation of the nine H. ducreyi isolates are listed in Table 1. The isolates belonged to either ribotypes 1, 2, or 4 (Table 1). The type strain (CIP 542) and 35000HP were susceptible to all antibiotics tested (Table 1). The isogenic LOS mutant, 35000HP-RSM2, was resistant to kanamycin, as expected. The six epidemiologically unrelated recent clinical isolates were variably resistant to penicillin, tetracycline, kanamycin, and chloramphenicol but were susceptible to azithromycin, erythromycin, ciprofloxacin, and ceftriaxone. The MICs for the clinical isolates were typical of those reported so far for most H. ducreyi strains in the 1990s, except for HMC 49, which was susceptible to penicillin and tetracycline (2, 6, 20).

Time course and dose-response assays.

Five H. ducreyi isolates were tested for susceptibility to PG-1 in classical killing assays. Poor growth in broth precluded testing of four of the recent clinical isolates. PG-1 rapidly killed the five isolates tested, causing ≥4-log₁₀-unit decrease within the first 15 min of the 2-h time course (data not shown).

H. ducreyi forms tight intracellular junctions that cause the organisms to clump in vivo and in vitro (10). We could not exclude the possibility that some of the apparent reduction in CFU due to treatment with PG-1 represented enhanced clumping of the bacteria by PG-1 rather than killing or inhibition of growth. To confirm that the bacteria were killed in the bactericidal assay, acetic acid and protegrin-treated 35000HP cells were stained for viability after 15 min of incubation with PG-1. Cells with intact membranes stained green (acetic acid control), while those with damaged membranes stained red (PG-1 treated) (Fig. 1). These data confirmed that PG-1 killed H. ducreyi.

FIG. 1.

Photomicrograph of H. ducreyi incubated with the acetic acid control (A) or with protegrin (B) For 15 min, stained with STYO 9 green and propidium iodide, and viewed with a fluorescein-rhodamine dual filter. Photographs were scanned on a Umax Astra 600S Scanner and assembled with Adobe Photoshop 3.0. Magnification, ×1,250.

Dose-response studies showed that concentrations of 6.3 to 12.5 μg of PG-1 per ml killed 2 log₁₀ CFU of the five isolates tested (Table 2). H. ducreyi 35000HP and its isogenic LOS mutant 35000HP-RSM2 were equally susceptible to PG-1 in this assay.

TABLE 2.

Dose-response study

Strain	Dose response at concn of PG-1 (μg/ml)a
	1.6	3.13	6.3	12.5
35000HP	0.1 ± 0.1	0.3 ± 0.2	1.2 ± 1.1	2.6 ± 1.3
35000HP-RSM2	0.07 ± 0.1	0.2 ± 0.3	1.1 ± 0.8	2.6 ± 1.9
CIP 542	0.3 ± 0.3	0.9 ± 0.2	2.1 ± 0.6	≥4b
HMC 48	0.6 ± 0.3	0.8 ± 0.1	2.5 ± 0.2	≥4
HMC 49	0.3 ± 0.3	0.8 ± 0.1	1 ± 0.5	2 ± 1.2

^aMeans ± standard deviations of log₁₀ CFU per milliliter in the acetic acid control wells divided by the log₁₀ CFU per milliliter in the protegrin-treated wells. 

^bMaximal detectabel killing was expressed as ≥4 log₁₀ CFU/ml, and maximal killing was obtained for all strains at 25 and 50 μg of PG-1 per ml. 

Radial diffusion assay.

To further exclude the possibility that PG-1 was clumping the organism, we tested the ability of PG-1 to inhibit the growth of all nine H. ducreyi isolates in a radial diffusion assay similar to that developed for N. gonorrhoeae (11). Plots of 35000HP and its isogenic LOS mutant 35000HP-RSM2 are shown in Fig. 2. The estimated minimal effective concentration of the peptide, which corresponds to the x intercept of the plots shown in Fig. 2 (8, 11, 16), ranged from 0.6 to 5 μg/ml (Table 1). In this assay, the minimal effective concentration of PG-1 for 35000HP-RSM2 was significantly lower than that for 35000HP (Table 1 and Fig. 2) (P = 0.004). Since entrapping the target cells in a gel precluded their agglutination by the peptide, these data confirm that PG-1 inhibits the growth of H. ducreyi.

FIG. 2.

Plots of radial diffusion assays for H. ducreyi 35000HP (squares) and 35000HP-RSM2 (diamonds).

DISCUSSION

Protegrins consist of a family of five highly homologous peptides of porcine origin (7, 11, 24, 25). The protegrins contain 16 to 18 amino acid residues and 4 invariant cysteines that form intramolecular disulfide bonds. In size and conformation, the protegrins most closely resemble tachyplesins, antimicrobial peptides found in hemocytes of horseshoe crabs (7). Protegrins also resemble defensins, antimicrobial peptides found in human neutrophils and rabbit heterophiles, but are considerably smaller than defensins and have a different spectrum of activity (7).

The sexually transmitted pathogens N. gonorrhoeae, C. trachomatis, herpes simplex virus type 2, and HIV-1 are all exquisitely susceptible to PG-1. Although HIV-1 and C. trachomatis are somewhat susceptible to defensins (17, 22), PG-1 is more potent than defensins in inactivating these organisms. Defensins have no activity against N. gonorrhoeae and lose activity against C. trachomatis in the presence of serum (11, 22). Thus, protegrins are better candidates than defensins for use as topical microbicides against STDs. Since PG-1 is as potent as the other protegrins in killing N. gonorrhoeae, we screened H. ducreyi for susceptibility to protegrins with synthetic PG-1.

In the liquid killing assay, both antibiotic-susceptible and -resistant strains of H. ducreyi were killed by PG-1. Concentrations of PG-1 ranging from 6.3 to 12.5 μg/ml killed 2 log₁₀ CFU of the five isolates tested. Bacteria with damaged membranes were visible after exposure to the peptide, confirming that the peptide killed the organism.

In the radial diffusion assay, the estimated minimal effective concentration of PG-1 for growth inhibition ranged from 0.6 to 5 μg/ml for the nine isolates tested. Similarly, the estimated minimal effective concentration of PG-1 for several N. gonorrhoeae strains ranges from 1 to 2 μg/ml (11). Protegrins permeabilize the outer membrane of gram-negative bacteria by binding to lipid A of lipopolysaccharide (unpublished data). The LOSs of H. ducreyi and N. gonorrhoeae are structurally and immunochemically similar, and the major glycoforms of both species terminate in N-acetyllactosamine capped by sialic acid (3, 9, 14). How protegrin interacts with LOS remains to be determined, but the similarity in the susceptibilities of H. ducreyi and N. gonorrhoeae to PG-1 is not surprising.

In the radial diffusion assay, transformants of N. gonorrhoeae FA19 that have truncated oligosaccharide structures (FA5101, containing a trisaccharide consisting of two KDOs and heptose; WS1, consisting of two KDOs) are more susceptible to PG-1 than the parent. An isogenic LOS mutant, 35000HP-RSM2, which contains a pentasaccharide consisting of glucose, a heptose trisaccharide core, and KDO, was more susceptible to PG-1 than its parent in the radial diffusion assay but not in the killing assay. Thus, the LOSs of H. ducreyi and N. gonorrhoeae may influence the susceptibility of these organisms to antimicrobial peptides in a similar manner.