Article Text


Expression of the cytolethal distending toxin in a geographically diverse collection of Haemophilus ducreyi clinical isolates
  1. K Kulkarni1,
  2. D A Lewis1,2,
  3. C A Ison1
  1. 1Department of Infectious Diseases and Microbiology, Faculty of Medicine, Imperial College, London W2 1PG, UK
  2. 2Department of Microbiology, University of Texas Southwestern Medical Center, 6000 Harry Hines Boulevard, Dallas, Texas 75235-9048, USA
  1. Correspondence to:
    Dr C A Ison, Department of Infectious Disease and Microbiology, Faculty of Medicine, Imperial College London, St Mary’s Campus, Norfolk Place, London W2 1PG, UK;


Objective: To screen a collection of isolates of Haemophilus ducreyi for expression of the cytolethal distending toxin (CDT).

Methods: 45 clinical isolates of H ducreyi were screened for cytotoxic activity by examining the effect of culture supernatants on Hela cells. Expression was confirmed using immunoblotting with CDT specific monoclonal antibodies and the presence of the cdt genes determined by amplification of the cdt genes in a multiplex polymerase chain assay.

Results: Of the 45 clinical isolates, six isolates from differing geographical origins did not demonstrate cytotoxic activity. Expression of CDT was also not detected in these six isolates using immunoblotting and the genes cdtA, cdtB, and cdtC were not amplified using PCR. The remaining isolates demonstrated cytotoxic activity, expressed the CDT proteins, and the presence of the cdt genes was confirmed.

Conclusions: CDT is considered a virulence factor of H ducreyi but was found to be absent in 13% of isolates from different geographical origins.

  • cytolethal distending toxin
  • Haemophilus ducreyi
  • chancroid

Statistics from

Chancroid is a sexually transmitted infection (STI), primarily of the tropics, caused by the Gram negative bacterium Haemophilus ducreyi. The organism causes ulcers on the genitalia, often with inguinal lymphadenitis.1 Untreated chancroid may persist for weeks or months, and can reoccur as there is no evidence of natural resistance, nor that past infection produces a protective immune response.2,3 Genital ulcerative diseases have been shown in numerous studies to be strongly associated with the acquisition of HIV.4

Many virulence factors have been reported1; of particular interest is a cytolethal distending toxin (CDT), one of two toxins produced by the organism, the other toxin being a bacterial cell associated haemolysin.5 The CDT is encoded by three closely linked genes—cdtA, cdtB, cdtC—which encode proteins with estimated molecular weights of 23 kDa, 29 kDa, and 19 kDa.6,7 The majority of H ducreyi isolates contain this gene cluster,8,9 indicating a significant role in pathogenesis.

The aim of this study was to screen a collection of H ducreyi isolates for cytotoxic activity, and to investigate the association with the cdt gene cluster.


Strains of H ducreyi

Forty five clinical isolates of H ducreyi were used in this study (table 1), and were from genital ulcers, rather than buboes, in patients seen in diverse geographical locations. These isolates were a random selection, whose identity had been confirmed and were known to be of diverse ribotypes.10,11 Two strains of H ducreyi, 35 00012 and 512,13 were used as positive and negative controls, respectively. H ducreyi CIP A77,14 which does not produce lesions in the temperature dependent rabbit model,15 was also included. Each isolate was retrieved from storage by inoculation onto a chocolate agar (CA) plate,16 and incubated at 36°C in a humidified atmosphere with 5% carbon dioxide. The purity was confirmed by the presence of characteristic Gram negative coccibacilli, and the ability of the colonies to be pushed across the medium. They were then subsequently subcultured and 14 hour cultures were used in these experiments.

Table 1

Cytotoxic activity of clinical isolates from various geographical locations

Preparation of culture supernatants

H ducreyi isolates were grown in a filter sterilised liquid medium.16 After incubation, purity was checked and the broths centrifuged, then the supernatant was ultracentrifuged at 46 000 rpm for 90 minutes at 4°C, filter sterilised, and stored at −70°C.

Cytotoxic assays

HeLa B cells (ECACC 85060701) were seeded at 3 × 104 cells per well in a 24 well tissue culture plate.17 After overnight incubation of the cells in 1 ml of tissue culture medium, an equal volume of filter sterilised bacteria culture supernatant was added to each well (three wells for each isolate), incubated for 3 hours before the fluid in each well was aspirated, and replaced with fresh tissue culture medium. The plates were incubated for 96 hours and then the test wells were compared with negative and positive control wells. Each well was scored as negative (confluent cells), positive (non-confluent cells), and photographed with an inverted phase contrast microscope. No graduations between positive and negative were noted after 96 hours of incubation.

Expression of CDT

Whole cells lysates were made from the pellets remaining after culture of H ducreyi strains in broth.18 The proteins were resolved by SDS-PAGE on 15% gels,19 with the outer two lanes loaded with rainbow markers (high molecular weight range, 14 300–220 000, and low molecular weight range, 2350–46 000, Amersham BioSciences). The proteins were transferred to nitrocellulose for western blot analysis20 and then probed with the H ducreyi CdtA reactive Mab 1G8, CdtB reactive Mab 20B2, and CdtC reactive Mab 8C9 kindly provided by Dr E Hansen.17 Anti-mouse IgG-HRP (Amersham Biosciences) was used to detect the antigen:antibody complex in each analysis and visualised using ECL detection reagents (Amersham Biosciences).

Amplification of cdt genes

Chromosomal DNA was prepared from cultures of each non-cytotoxin producing H ducreyi and a selection of cytotoxin producing isolates harvested from chocolate agar plates incubated for 14 hours at 36°C (Qiamp kit, Qiagen).

A multiplex polymerase chain reaction (PCR) was constructed to determine the existence of any or all of the cdt genes. The sequences for each cdt primer pair were; cdtA, forward (cdt-5) 5’-128AGG ATG GAT CTA AGG AGA G,146 3’reverse (HB13–2) 5’-603GTG CTA ATT TGT CCA GAC620-3’(PCR product 475 bp); cdtB, forward (cdtBC) 5’-1179 GCA AAC CGA GTG AAC TTA G1198-3’ reverse (cdt-1) 5’-1481TAT TTT CAC TCA CTG CGG1498-3’ (PCR product 320 bp); cdtC, forward (cdt-9) 5’-1984ATG TTT TGC TTT CCT GGG2001-3’ reverse (cdt-11) 5’-2237ACC CTG ATT TCT TCG CAC2254-3’ (PCR product 269 bp). Oligonucleotide primers to detect 16S ribosomal RNA of eubacteria were also added as a control, forward, 5’-509AAC T(C/A)A CGT GCC AGC AGC CGC GGT A533-3’ reverse 5’1517-AAG GAG GTG ATC CA(G/A) CCG CA(G/C) (G/C)TT C1541-3’.

The PCR mix used consisted of 5 μl of 10X PCR buffer (Gibco BRL), 3 μl 50 mM MgCl2 (Gibco BRL), 2 μl 10 mM dNTPs (Boehringer Mannheim), 1 μl of each cdt primers (Sigma-Genosys) at a concentration of 50 pmol/μl, 2 μl of the 16S ribosomal RNA primers (Sigma-Genosys) at a concentration of 50 pmol per μl, 28 μl of dH2O, 2 μl of target DNA, and 2 units Taq DNA polymerase (Gibco BRL). The PCR was performed in Robocycler Gradient 40 temperature cycler (Strategene), and the mix incubated at 94°C for 5 minutes to activate the Taq polymerase, followed by 25 cycles of 94°C for 1 minute, 51°C for 1 minute, and 72°C for 3 minutes. The final cycle was 72°C extension step for 5 minutes. The resulting PCR products were analysed using DNA 7500 chips on an Agilent bioanalyser (Agilent Technologies).


Cytotoxic assay results

In comparison with the control wells bacterial culture supernatants from 39 (87%) of the isolates, as exemplified by H ducreyi isolate D099 (fig 1A), exhibited HeLa cell cytotoxicity (table 1). After 96 hours, marked distension can be seen in the HeLa cells, continuing to the point where they disintegrate. At this time the cells detached from the substratum, destroying the cell monolayer. Supernatants obtained from six clinical isolates, including H ducreyi D033 (fig 1B), produced no cytotoxic activity. The monolayer of HeLa cells was not destroyed, but was intact growing across the base of the well to become confluent (table 1). H ducreyi CIP A77, known to be virulent in the animal model, also exhibited cytotoxicity.

Figure 1

Effect of culture supernatant on a HeLa cell monlayer from (A) positive strain, D099 and (B) a negative strain, D033.

Expression of CDT

In all isolates positive in the cytotoxic assay, a band corresponding to the putative form of CdtA of molecular mass of 23 kDa was detected (fig 2A). It was not present in isolates which were negative in the cytotoxic assay. Similar results were found for CdtB (fig 2B) and CdtC (fig 2C) with molecular masses of 29 kDa and 19 kDA respectively.

Figure 2

Expression of the proteins (A) CdtA, (B) CdtB, and (C) CdtC by isolates exhibiting no cytotoxic activty (−) and a selection of isolates showing cytotoxic activity (+). Strain 35 000 is the positive control and strain 512 is the negative control, D denotes a clinical isolate.

cdt genes

Amplification of cdt genes and control 16S rRNA in a multiplex PCR showed that isolates that were positive in the cytotoxic assay and expressed the protein each showed the expected PCR products of 475 bp (cdtA), 320 bp (cdtB), and 269 bp (cdtC) (fig 3). In each of the isolates that were negative in the cytotoxic assay and did not express the protein, only the 16S rRNA was amplified (fig 3), implying that these strains do not contain the cdt gene cluster.

Figure 3

Amplification of the cdtA, cdtB, and cdtC genes in isolates exhibiting no cytotoxic activty (−) and a selection of isolates showing cytotoxic activity (+). Strain 35 000 is the positive control and strain 512 is the negative control, D denotes a clinical isolate.


In this study six out of the 45 isolates tested for cytotoxic activity were unable to induce cell death in HeLa cells, did not express any of the Cdt proteins, and the cdt gene cluster was not detected. The presence of the cdt gene cluster in 87% of isolates is similar to findings in other collections of isolates of 89%8 and 82%.9 The isolates in this study were isolated in different geographical locations but it is not possible to determine any particular association with country of isolation or source of infection because this was a random collection of isolates rather than a sentinel study.

After the original discovery of CDT by Johnson and Lior,21 Purvén et al8 described the subsequent characterisation of CDT effects in in vitro cell culture, and the soluble nature of H ducreyi CDT. In other bacterial species cytotoxic activity is known to be due to the expression of cdtABC gene cluster,22–,24 which encodes the cytolethal distending toxin. The effects of this toxin are not to be confused with that of the H ducreyi haemolysin, which is lethal for human foreskin fibroblasts (HFFs) but does not affect epithelial cells (HeLa cells),25,26 whereas CDT is cytotoxic to many different cell lines.6,27–,29 A haemolysin deficient30 CDT deficient mutant and a mutant deficient in both haemolysin and CDT31 have been tested in the human model for experimental chancroid and have no effect on pustule formation, indicating neither has a direct role at this stage of the infection.

The role of CDT and its purpose in the course of chancroid remains unclear. It is known that CDT causes sensitive eukaryotic cells to be halted in the G2 phase of the cell cycle.27 These cells, unable to proliferate, distend up to five times their original size before disintegrating (fig 1). The same phenotypic effects on the cell cycle have been detected in various studies with HIV infected cells, or cells transfected with vpr, an HIV regulatory gene.32,33 It has been suggested that the CDT holotoxin has intrinsic DNase activity that is associated with the CdtB polypeptide.34 It may be this DNase activity that is responsible for the cell cycle arrest, which CDT is capable of inducing.

In this study approximately 13% of the strains were non-cytotoxin producing, it is unlikely that these have arisen spontaneously. It may be that H ducreyi originally did not possess the ability to produce CDT and the cdt gene sequence may have been acquired from a heterologous species. This is supported by the presence of transposase relics and one apparently intact transposase gene within 3 kb on either side of the H ducreyi cdt genes.7

CDT may be involved in genital ulceration in different ways. In the epidermis, it is able to cause the irreversible block of cell proliferation, consistent with the generation and slow healing of ulcers. After progressing through the epidermis, the toxin could act upon the fibroblasts in the dermis. H ducreyi CDT can destroy both keratinocytes and T lymphocytes,35 suggesting that CDT may interfere with T cell responses to H ducreyi by apoptosis thus creating an environment that supports continued growth of the bacterium and facilitates enlargement and/or persistence of the ulcer. In this way it can be seen that H ducreyi CDT has potential to be a major virulence factor for the organism. Despite their in vitro effects these non-cytotoxin producing clinical isolates have been isolated from symptomatic patients, confirming the multivirulence determinant nature of the organism. The effect of haemolysin, the immune status of the individual and co-infection of other STIs, and the presence of a foreskin may all contribute to the ulceration in these patients, thereby compensating for the lack of CDT production.


We are grateful to Dr Eric Hansen, Department of Microbiology at the University of Texas Southwestern Medical Center, Dallas, USA, for providing both the monoclonal antibodies and H ducreyi isolates 35 000 and 512. We also thank Natasha Anwar and Colin Sandiford for technical assistance and Odile Harrison for the 16S rRNA primers. DAL was funded by a Wellcome training fellowship in clinical tropical medicine (reference number 049246/Z/96).

Conflicts of interest: none.

KK undertook this study as part of an intercalated BSc in infection and immunity at the Faculty of Medicine, Imperial College of Science, Technology and Medicine and was awarded the MSSVD Undergraduate Science prize; KK was responsible for the technical work and preparation of the manuscript; DAL and CAI acted as supervisors and provided advice and critical review of the manuscript.


View Abstract

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.