INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

The genital ulcer disease chancroid, caused by the bacterium Haemophilus ducreyi, is one of the most prevalent sexually transmitted diseases and is a major cause of morbidity in the resource-poor countries of Africa, Asia, and Latin America (50). In the pre-antibiotic era, chancroid ulcers were noted to heal at slow rates and often resolution was incomplete (38, 50). Bacterial factors, acting either directly or indirectly, may be responsible for this observed chronicity of H. ducreyi genital ulceration. In particular, it is possible that one or both of the two toxins known to be elaborated by H. ducreyi could be involved in this effect. The hemolysin HhdA, which has been described as being bacterial cell-associated, kills human foreskin fibroblasts (HFF) in vitro (1, 28). The other H. ducreyi cytotoxin, a member of the family of toxins known as cytolethal distending toxins (CDTs) (33), has little effect on HFF cells but readily kills other human cells, including HeLa cells in vitro (37, 44), and has been shown to cause Jurkat T cells to undergo apoptosis (14).

The vast majority of H. ducreyi strains (35) produce CDT that can be readily detected in culture supernatant fluid. The H. ducreyi CDT, first described phenotypically by Purven and Lagergard (37), is chromosomally encoded by three adjacent genes, cdtA, cdtB, and cdtC, which appear to be transcriptionally linked (8). Similar CDTs are expressed by a number of different enteric organisms, including Escherichia coli (18, 31, 40), Shigella species (17, 27), and Campylobacter species (19, 32), as well as by Actinobacillus actinomycetemcomitans (25, 46). CDT activity was originally characterized as causing relatively slow morphological changes in cultured epithelial cells, including progressive cellular distention and death within 96 to 120 h (18). CDT exerts its effect through cell cycle arrest in the G₂ phase of growth (7, 9, 30, 41, 46, 52). Most recently, it was reported that CDT has intrinsic DNAse activity associated with the CdtB polypeptide (11, 12, 22).

The exact composition of the active form of CDT is as yet unknown. All three gene products (i.e., CdtA, CdtB, and CdtC) appear to be necessary for the expression of E. coli CDT activity in culture supernatant fluid (31, 40). There is also evidence to support the involvement of the H. ducreyi cdtC gene product in the expression of cytotoxic activity because a monoclonal antibody (MAb) to the H. ducreyi CdtC protein has been shown to neutralize H. ducreyi CDT activity in vitro (8, 25), although it is unlikely that CdtC itself has toxic activity like that attributed to CdtB (41).

In the present study, we report the construction and characterization of isogenic cdtA and cdtB mutants of H. ducreyi 35000. The phenotypes of these two new mutants and that of a H. ducreyi cdtC mutant (44) were assessed through the use of in vitro cytotoxicity assays with HeLa cells and in the temperature-dependent rabbit model for experimental chancroid. Cytotoxic activity assays of whole-cell sonicates and periplasmic extracts prepared from these three isogenic mutants revealed that these preparations from the cdtA mutant still possessed some CDT activity, whereas the same preparations from the cdtB and cdtC mutants were devoid of CDT activity.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Bacterial strains and culture conditions. The bacterial strains and plasmids used in this study are listed in Table 1. All H. ducreyi strains were routinely cultivated on chocolate agar (CA) plates containing 1% (vol/vol) Iso VitaleX (BBL Microbiological Systems, Cockeysville, Md.) in a humidified atmosphere of 95% air-5% CO₂ at 33°C as previously described (23). Broth cultures of H. ducreyi were grown overnight (16 h) in modified Columbia broth (51) for the preparation of whole-cell sonicates, periplasmic extracts, and culture supernatant fluids. Cells of E. coli strains were grown on Luria-Bertani medium (39) at 37°C. For antimicrobial supplementation, kanamycin was used at 30 µg/ml (for both E. coli and H. ducreyi), tetracycline was used at 15 µg/ml (for E. coli), and chloramphenicol was used at 0.5 µg/ml (for H. ducreyi) and 15 µg/ml (for E. coli).

MAbs. The H. ducreyi CdtA-reactive MAb 1G8, the CdtB-reactive MAb 20B2, and the CdtC-reactive MAb 8C9 have been described (44).

Recombinant DNA techniques. Standard techniques, including restriction enzyme digests, ligation, transformation, and plasmid purification, have been described elsewhere (4, 39). An aphA-3 gene (kan1) lacking both its own promoter and transcriptional terminators was utilized for the construction of nonpolar insertion mutations in the H. ducreyi cdtA and cdtB genes. This aphA-3 gene is preceded by translational stop codons in all three reading frames and is immediately followed by a consensus ribosome binding site and a start codon. The use of appropriate restriction sites allows placement of the start codon (at the 3' end of this cassette) in the open reading frame of the stop codon of the mutated gene so that translation of the remaining 3' portion of the mutated gene will allow translational coupling, if any, with a downstream gene(s). This aphA-3 gene was excised from pUC18K by digestion with SmaI (26). PCR was performed according to manufacturers' instructions with either Taq DNA polymerase (Promega, Madison, Wis.) or Pfu DNA polymerase (Stratagene, La Jolla, Calif.). Boiled bacterial cell preparations (16) or purified H. ducreyi chromosomal DNA (5) were used as templates for PCR.

Nucleotide sequence analysis. DNA sequencing was performed using a model 373A automated DNA sequencer (Applied Biosystems Inc., Foster City, Calif.). All PCR products used for complementation analyses were sequenced to confirm the absence of PCR-derived mutations.

RT-PCR detection of H. ducreyi mRNA from rabbit lesions. A lesion from a rabbit injected intradermally 48 h previously with 10⁵ CFU of H. ducreyi 35000 was excised and macerated in 1 ml of Ultraspec RNA isolation reagent (BiotecX Laboratories, Inc., Houston, Tex.) by using a glass homogenizer and then stored at 70°C until processed. After suspension of the purified RNA pellet in 50 µl of diethyl pyrocarbonate-H₂O, the RNA sample was treated with 10 U of DNase at 37°C for 30 min to degrade any residual DNA using the Message Clean^R Kit (GenHunter Corporation, Nashville, Tenn.).

Construction of isogenic H. ducreyi cdtA and cdtB mutants. The cdtA gene in the cdtABC gene cluster in pJL300 (44) was inactivated by insertion of the kan1 cartridge into the RsrII site (Fig. 1) (26) to produce pJL301. Inactivation of the cdtB gene was accomplished by inserting the same kan1 cartridge into the MscI site within the cdtB gene in pJL300; the resultant construct was designated pJL302. Selection for kanamycin resistance followed by nucleotide sequence analysis confirmed that the cartridge was inserted in the proper orientation in each open reading frame. Plasmids pJL301 and pJL302 were purified from E. coli HB101 cells by using the Wizard Plus minipreps purification system (Promega), linearized by digestion with AatII, and used to electroporate H. ducreyi 35000. Transformants were selected on GC-heme plates supplemented with kanamycin (45). H. ducreyi transformants with potential mutations in the cdtA gene were identified by PCR-based amplification of chromosomal DNA from single colonies by using the oligonucleotide primers P9 (5'-CCTTGTAGATTATTCACCGT-3') and P10 (5'-AAAGTTTGATGGTTCAGACGG-3') and by the loss of reactivity with the H. ducreyi CdtA-reactive MAb 1G8 using Western blot analysis. Similarly, H. ducreyi transformants with potential mutations in the cdtB gene were identified by PCR-based amplification of chromosomal DNA from single colonies using oligonucleotide primers P11 (5'-GCAAACCGAGTGAACTTAG-3') and P12 (5'-TATTTTCACTCACTGCGG-3') and by loss of reactivity with the H. ducreyi CdtB-reactive MAb 20B2. The isogenic H. ducreyi 35000 cdtC mutant used in this study has been described (44).

View larger version (18K): [in this window] [in a new window]   FIG. 1.   Partial restriction map of the H. ducreyi 35000 chromosomal DNA insert in pJL300 and related plasmids. Restriction sites in parentheses indicate vector cloning sites. The arrows indicate the direction of transcription. Plasmid pJL300 was derived from the pBR322 vector; both it and pJL303 have been described (44). The kan1 cartridge was blunt-ended and ligated into the similarly treated RsrII site in cdtA to construct pJL301 and into the similarly treated MscI site of cdtB to produce pJL302. Plasmid pDL20-A is a modified pBR322 vector containing the 1.1-kb PstI- and EcoRI-digested PCR product with the cdtA gene. Plasmid pDL10-B is a modified pACYC184 vector containing the 1.2-kb SalI- and BamHI-digested PCR product with the cdtB gene. Plasmid pJL300-C is pLS88 containing the cdtC gene (44).

Construction of a plasmid containing the H. ducreyi cdtA gene for complementation studies in E. coli. A 1.1-kb DNA fragment containing the cdtA gene was amplified from H. ducreyi 35000 chromosomal DNA by PCR using Pfu DNA polymerase and the oligonucleotide primers P13 (5'-TACTGCAGTTCCCATACGCCAGGATAG-3') and P14 (5'-TAGAATTCATAACATCACACAGAAAACCACAC-5'); the underlined sequences indicate PstI and EcoRI sites, respectively. The PCR product was digested with PstI and EcoRI before ligation into the 3.6-kb PstI-EcoRI fragment of the plasmid vector pBR322. The ligation mixture was used to transform E. coli DH5; tetracycline-resistant transformants were screened for expression of CdtA by Western blot analysis using MAb 1G8. Plasmid pDL20-A containing the H. ducreyi cdtA gene was obtained from one of these transformants. In addition, E. coli DH5 was transformed with the 3.6-kb pBR322-derived vector pDL19-2 (Table 1).

Construction of a plasmid containing the H. ducreyi cdtB gene for complementation studies in E. coli. A 1.2-kb DNA fragment containing the cdtB gene was amplified from H. ducreyi 35000 chromosomal DNA as described above using the oligonucleotide primers P15 (5'-TAGGATCCTGGTGCGGTTGTCATTAAAAG-3') and P16 (5'-ATTAGTCGACGAGGAGGTGATAACTCTACATCAGG-3'); the underlined sequences indicate BamHI and SalI sites, respectively. The PCR product was digested with BamHI and SalI before ligation into the 4-kb BamHI-SalI fragment of the plasmid vector pACYC184 (6). The ligation mixture was used to transform E. coli DH5; chloramphenicol-resistant transformants were screened for expression of CdtB by Western blot analysis using MAb 20B2. Plasmid pDL10-B was obtained from one of these transformants (Table 1). Plasmid pACYC184 was digested with both BamHI and SalI and, after blunt-ending, was self-ligated to form the vector pDL16-1.

Preparation of bacterial extracts for cytotoxicity testing. E. coli culture supernatant fluid preparations were prepared as previously described (44). To obtain H. ducreyi culture supernatant fluid preparations, 16-h overnight broth cultures were subjected to centrifugation at 7,600 × g for 20 min to collect the bacterial cells which were then used to prepare whole-cell sonicates and periplasmic extracts as described below. The resultant supernatant was sterilized by filtration through a cellulose acetate filter (0.2-µm pore size; Nalgene, Rochester, N.Y.) and then subjected to centrifugation at 219,000 × g for 1.5 h. The final supernatant fluid from this last centrifugation step was filter sterilized again and used as the culture supernatant fluid for cytotoxicity testing. To prepare whole-cell sonicates, the bacterial cells collected by centrifugation from the broth culture were washed once with pH 7.3 phosphate-buffered saline and collected again by centrifugation. The cell pellet was suspended in 20 mM Tris-HCl (pH 8.0) containing 20% (wt/vol) sucrose, using 5 ml of this solution per gram (wet weight) of cells. This suspension was sonicated five times (1-min sonication followed by a 1-min cooling period each time) using a Branson model 450 sonifier (Branson Sonic Power Co., Danbury, Conn.) and a tapered microtip on power setting no. 10 and at 60% duty. The whole-cell sonicate was subjected to centrifugation at 27,000 × g for 20 min to remove unbroken cells and debris. The resultant whole-cell sonicate was sterilized by filtration through a cellulose acetate filter (0.2-µm pore size) and stored at 4°C overnight prior to performing the cytotoxicity assay. Periplasmic extracts were prepared as previously described (23), filter sterilized, and stored at 4°C until used to perform cytotoxicity assays. To confirm the presence of periplasmic proteins in the periplasmic extract, MAb 3F1, which is specific for the H. ducreyi periplasmic ZnuA protein (23), was used in Western blot analysis. In addition, periplasmic extracts and culture supernatant fluids were tested for the presence of cell envelopes by using MAb 3F12, specific for the H. ducreyi major outer membrane protein (MOMP) (21). These preparations were also tested for the presence of cytosolic contamination by assaying the cytoplasmic enzyme glucose-6-phosphate dehydrogenase (24).

Cytotoxicity assays. Cytotoxicity assays for CDT activity used HeLa cells (ATCC CCL-2) grown in supplemented Dulbecco's modified Eagle's medium (GIBCO-BRL, Gaithersburg, Md.) in 24-well plates as previously described (44). After incubation of freshly seeded HeLa cells (3 × 10⁴ cells/well) in 0.5 ml of tissue culture medium overnight, either a sample of bacterial culture supernatant fluid (0.5 ml) or a sample of H. ducreyi periplasmic extract or whole-cell sonicate (50 µg of protein in 0.5 ml of tissue culture medium) was added to the HeLa cells for 3 h before removal of the fluid and replacement with fresh tissue culture medium. Serial twofold dilutions of these three preparations were made in tissue culture medium to determine CDT titers. The plates were incubated for 96 h at 37°C in an atmosphere of 95% air-5% CO₂. At the 96 h time point, the wells were either stained with Giemsa and photographed or the CellTiter 96_Aqueous One Solution Cell Proliferation Assay (Promega) was used to determine the extent of killing (i.e., which wells still had viable cells). The CDT titer was defined as the greatest dilution which did not produce detectable killing of the HeLa cells.

Preparation of soluble cell proteins and Sarkosyl-insoluble cell fractions. CA plate-grown H. ducreyi cells were subjected to sonication as described above. The sonicate was centrifuged at 3,000 × g for 10 min to remove whole cells and gross debris; the supernatant fluid from this centrifugation step was then subjected to centrifugation at 40,000 × g for 1 h to collect cell envelopes. The supernatant fluid from this second centrifugation step was subjected to ultracentrifugation at 156,000 × g for 2 h to remove any membrane fragments; the final supernatant fluid was designated as the soluble cell fraction. The cell envelopes were washed twice with phosphate-buffered saline and then treated with Sarkosyl as previously described (13) to obtain the Sarkosyl-insoluble fraction.

Sodium dodecylsulfate-polyacrylamide gel electrophoresis and Western blot (immunoblot) methods. Proteins present in various cell fractions and concentrated culture supernatant fluids were resolved by sodium dodecyl sulfate-polyacrylamide gel electrophoresis and transferred to nitrocellulose paper for Western blot analysis with MAbs as previously described (44).

Virulence testing. The relative virulence of the three H. ducreyi cdt mutants used in this study was determined in a blinded manner using the temperature-dependent rabbit model for experimental chancroid (34). Lesion characteristics on days 2, 4, and 7 postinfection were scored with the following numeric values: 0 = no change, 1 = erythema, 2 = induration, 3 = nodule, 4 = pustule or necrosis. On day 7 postinfection, material excised from lesions caused by injection of 10⁵ CFU was cultured on CA plates. Statistical analyses were performed as previously described (2, 45).

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Complementation studies with E. coli. In a preliminary effort to evaluate the contribution of the CdtA, CdtB, and CdtC proteins to the cytotoxic activity of H. ducreyi CDT, E. coli was used as the host for various combinations of the H. ducreyi cdtA, cdtB, and cdtC genes carried individually on plasmids. Two new plasmids with H. ducreyi cdtA (i.e., pDL20-A) and cdtB (i.e., pDL10-B) gene inserts were constructed for this purpose (Table 1 and Fig. 1). Plasmid pJL300-C containing a functional H. ducreyi cdtC gene (44) was also used (Table 1). Western blot analysis of whole-cell lysates of the various recombinant E. coli strains with CdtA-reactive, CdtB-reactive, and CdtC-reactive MAbs confirmed that each construct expressed the expected cdt gene product (data not shown). Assays for CDT activity were performed using culture supernatant fluids obtained from overnight cultures of strains possessing each gene individually as well as all possible double or triple combinations of the three cdt genes and their relevant vector controls. In addition, E. coli strains DH5(pBR322) and DH5(pJL300) (Table 1) were used as negative and positive controls, respectively, in the cytotoxicity assays.

Construction of isogenic H. ducreyi cdtA and cdtB mutants. To construct an isogenic H. ducreyi cdtA mutant, pJL301 (Fig. 1) was linearized and used to electroporate H. ducreyi 35000. A kanamycin-resistant transformant which failed to react with the H. ducreyi CdtA-reactive MAb 1G8 was selected for further study and designated strain 35000.301. Western blot analysis revealed that this cdtA mutant expressed no detectable CdtA protein (Fig. 2A, lane 2) and readily detectable levels of both CdtB (Fig. 2B, lane 2) and CdtC (Fig. 2C, lane 2).

View larger version (34K): [in this window] [in a new window]   FIG. 2.   Western blot-based detection of H. ducreyi CdtA, CdtB, and CdtC proteins expressed by wild-type and mutant H. ducreyi strains. Whole-cell sonicates (lanes 1 to 4) and periplasmic extracts (lanes 5 to 8) of these four strains were probed with the H. ducreyi CdtA-reactive MAb 1G8 (A), the H. ducreyi CdtB-reactive MAb 20B2 (B), the H. ducreyi CdtC-reactive MAb 8C9 (C), the ZnuA-reactive MAb 3F1 (D), and the MOMP-reactive MAb 3F12 (E). Equivalent amounts of whole-cell sonicate and periplasmic extract were loaded into each lane. Lanes: 1 and 5, wild-type H. ducreyi 35000; 2 and 6, H. ducreyi cdtA mutant 35000.301; 3 and 7, H. ducreyi cdtB mutant 35000.302; 4 and 8, H. ducreyi cdtC mutant 35000.303. Molecular size markers (in kilodaltons) are indicated on the right.

Cytotoxic activity of the wild-type and mutant H. ducreyi strains. Initially, culture supernatant fluids from the wild-type strain 35000, the cdtA mutant 35000.301, the cdtB mutant 35000.302, and the cdtC mutant 35000.303 (44) were tested for HeLa cell cytotoxicity. These culture supernatant fluids were free of cytosolic contamination as determined by the absence of detectable glucose-6-phosphate dehydrogenase activity (data not shown) and did not contain either periplasmic proteins or outer membranes, as evidenced by their lack of reactivity with the ZnuA-specific MAb 3F1 and the MOMP-specific MAb 3F12, respectively (data not shown). Cytotoxicity was observed with culture supernatant fluids from the wild-type strain only (Fig. 3A); culture supernatant fluids from the three isogenic mutants, 35000.301, 35000.302, and 35000.303, all failed to exhibit cytotoxic activity. When CdtA-, CdtB-, and CdtC-directed MAbs were used in Western blot analysis to probe culture supernatant fluid obtained from the wild-type parent strain, none of the Cdt proteins could be detected (data not shown). Fortyfold concentrations of these supernatant fluids still did not allow detection of Cdt proteins by Western blot analysis (data not shown).

View larger version (20K): [in this window] [in a new window]   FIG. 3.   Killing of HeLa cells by CDT in subcellular fractions and culture supernatant fluids from wild-type and mutant strains of H. ducreyi. Cytotoxic activity was assessed in culture supernatant fluids (A), periplasmic extracts (B), and whole-cell sonicates (C) obtained from the wild-type H. ducreyi strain 35000, the cdtA mutant 35000.301, the cdtB mutant 35000.302, and the cdtC mutant 35000.303. HeLa cells were exposed to filter-sterilized culture supernatant fluids, periplasmic extracts, or whole-cell sonicates as described in Materials and Methods. After 96 h, wells in which viable cells were present were detected by the use of the tetrazolium reduction-based assay described in Materials and Methods. An optical density at 492 nm (OD492) greater than 1.8 indicates wells in which all of the HeLa cells were viable.

Cytotoxicity of subcellular fractions from wild-type and mutant strains. Both periplasmic extracts and whole-cell sonicates from the wild-type strain and the cdtA, cdtB, and cdtC mutants were tested for their ability to kill HeLa cells. The periplasmic extracts were free of cytosolic contamination (data not shown) and outer membrane fragments (Fig. 2E, lanes 5 to 8) as determined above and did contain the periplasmic protein ZnuA (Fig. 2D, lanes 5 to 8). The relevant Cdt proteins could be detected readily in the whole-cell sonicates (Fig. 2A to C, lanes 1 to 4). However, in most cases, Cdt proteins could not be detected by Western blot analysis in the periplasmic extracts (Fig. 2A to C, lanes 5 to 8), even with the wild-type strain. Both whole-cell sonicates (Fig. 3C) and periplasmic extracts (Fig. 3B) from wild-type strain 35000 caused marked killing of the HeLa cells, whereas the same preparations from the cdtA mutant 35000.301, the cdtB mutant 35000.302, and the cdtC mutant 35000.303 appeared to lack killing activity as determined by a tetrazolium dye reduction test (described in Materials and Methods). However, microscopy of the HeLa monolayers exposed to the most concentrated samples (i.e., 50 µg of protein in 0.5 ml) of whole-cell sonicate from the cdtA mutant 35000.301 (Fig. 4B) revealed the presence of many distended or rounded HeLa cells. These distended cells were not present in the monolayers treated with whole-cell sonicates from the cdtB mutant (Fig. 4C) and the cdtC mutant (Fig. 4D). No viable HeLa cells were present in monolayers treated with the whole-cell sonicate from the wild-type parent strain (Fig. 4A). The periplasmic extract from the cdtA mutant also caused HeLa cell rounding and distention whereas the periplasmic extracts from the cdtB and cdtC mutants did not affect these cells (data not shown).

View larger version (52K): [in this window] [in a new window]   FIG. 4.   Cytopathic effect of CDT from wild-type and mutant strains of H. ducreyi. HeLa cells were exposed to whole-cell sonicates from the wild-type strain 35000 (A), the H. ducreyi cdtA mutant 35000.301 (B), the H. ducreyi cdtB mutant 35000.302 (C), and the H. ducreyi cdtC mutant 35000.303 (D), and 96 h later the wells were photographed at an original magnification of ×40. (E) HeLa cells exposed to the buffer used for the preparation of the whole-cell sonicates were used as the negative control.

View larger version (36K): [in this window] [in a new window]   FIG. 5.   Detection of Cdt proteins in soluble and insoluble fractions from H. ducreyi cells. Whole-cell sonicates were prepared from the wild-type strain 35000 as described in Materials and Methods by using centrifugation at 3,000 × g to remove whole cells and gross debris. Equivalent amounts of whole-cell sonicate (lane 1), soluble cell contents (lane 2), cell envelopes (lane 3), and Sarkosyl-insoluble material (lane 4) were probed by Western blot analysis using the CdtA-reactive MAb 1G8 (A), the CdtB-reactive MAb 20B2 (B), the CdtC-reactive MAb 8C9 (C), and the MOMP-reactive MAb 3F12 (D). Molecular size position markers (in kilodaltons) are on the right.

Virulence testing. After confirming that the three isogenic mutants 35000.301, 35000.302, and 35000.303 each grew at a rate indistinguishable from that of the wild-type strain 35000 in broth (data not shown), all four strains were tested in the temperature-dependent rabbit model. The cdtA and cdtB mutants both proved to be as virulent as the wild-type parent strain with regard to lesion production (Table 2). As seen previously (44), the cdtC mutant 35000.303 also was no less virulent than the wild-type strain (Table 2). In addition, viable H. ducreyi organisms were isolated from the lesions resulting from an injection of 10⁵ CFU for all four strains (data not shown). The numbers of CFUs of each of the four strains recovered from these lesions were similar (data not shown).

Detection of cdtA, cdtB, and cdtC mRNA produced in vivo. The fact that the mutants lacking CDT activity were as virulent as the wild-type strain in the rabbit model prompted us to confirm that the cdtABC genes were actually transcribed in the cells of the wild-type parent strain growing in vivo. RNA was prepared from pustular lesion material obtained 48 h after infection of rabbits with the wild-type strain, 35000. RT-PCR confirmed the presence of mRNA transcribed from the cdtA, cdtB, and cdtC genes. Specific primers amplified a 132-bp cdtA-encoded product, a 460-bp cdtB-encoded product, and a 271-bp cdtC-encoded product from the RNA (Fig. 6, lane 3). A 355-bp pal-encoded product was amplified from the RNA as a positive control (Fig. 6, lane 3); the H. ducreyi pal gene encodes an unprocessed 18-kDa lipoprotein (43). H. ducreyi 35000 chromosomal DNA was used as a template positive control (Fig. 6, lane 1). No PCR product was seen in a reaction devoid of template (Fig. 6, lane 2) or in a reaction in which the RT step was omitted (Fig. 6, lane 4).

View larger version (87K): [in this window] [in a new window]   FIG. 6.   Agarose gel electrophoresis of H. ducreyi multiplex RT-PCR products. The sizes of the predicted RT-PCR products were as follows: pal, 355 bp; cdtA, 132 bp; cdtB, 460 bp; cdtC, 271 bp. Templates included the following: lane 1, 100 ng of H. ducreyi 35000 genomic DNA (positive control); lane 2, no template (negative control); lanes 3 and 4, 2 µg of RNA isolated from H. ducreyi-infected rabbit lesions. The reaction mixture loaded in lane 4 was not subjected to the reverse transcription step of the RT-PCR protocol and served as a control to detect DNA contamination of the RNA template. Lane M, DNA size markers (in base pairs).

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

Johnson and Lior (18) originally described CDT activity as a toxic factor produced by certain E. coli strains which would distend and kill CHO cells in vitro. CDT production in several unrelated pathogenic bacteria was subsequently demonstrated (33). CDT can affect a number of different cell types in vitro, including keratinocytes (9, 44), HeLa cells, HEp-2 cells, Vero cells (20), Caco-2 cells (52), hamster lung fibroblasts (9), and both CD4⁺ and CD8⁺ human T cells (42). While new data have appeared recently concerning the mechanism of action of CDT (7, 9, 30, 46, 52), there is little information about the composition of the CDT holotoxin.

To date, there have been only two reports describing the purification of a CDT. Working with H. ducreyi, Lagergard and colleagues (36) used a MAb specific for CdtC in an immunoaffinity-based purification method and concluded that H. ducreyi CDT consisted of only CdtC. In contrast, Shenker et al. (42) purified an immunosuppressive factor from A. actinomycetemcomitans cells that killed human T cells and HeLa cells and indicated that this purified toxin was comprised of CdtB. (The protein products of the A. actinomycetemcomitans cdtABC genes are 92 to 97% identical to those of the H. ducreyi cdtABC genes [25].) In neither case was the purity of the protein preparation demonstrated conclusively; in particular, the absence of the other two cdt gene products was not proven.

Data derived from the use of individually cloned H. ducreyi cdtA, cdtB, and cdtC genes in an E. coli background demonstrated that CDT activity in culture supernatant fluid was detectable only when all three cdt gene products were expressed. This result reinforces results obtained in earlier studies of the cdtABC gene clusters from two different E. coli strains (31, 40). Similarly, culture supernatant fluids from isogenic H. ducreyi cdtA, cdtB, and cdtC mutants all failed to kill HeLa cells, whereas culture supernatant fluid from the wild-type parent strain caused extensive cell killing (Fig. 3). In contrast to the results obtained with culture supernatant fluids, both whole-cell sonicates and periplasmic extracts prepared from the H. ducreyi cdtA mutant produced visible cytopathic effects in HeLa cells, with cell distention being most prominent. It should be noted that cell distention caused by CDT precedes killing (18, 20), such that the cell distention effect of the whole-cell sonicate (Fig. 4B) and periplasmic extract of the cdtA mutant likely reflects the presence of either a small quantity of active CDT or a partially inactive CDT. More importantly, the whole-cell sonicates and periplasmic extracts prepared from the H. ducreyi cdtB and cdtC mutants showed no evidence of cytotoxic activity in the HeLa cell assay.

The composition of active H. ducreyi CDT remains to be determined. The results described above indicate that mutations in cdtB and cdtC eliminate CDT activity completely, whereas some very limited CDT activity is expressed by the cdtA mutant. These results, taken together with the facts that MAbs to CdtC can neutralize CDT activity (8, 25, 36) and that the CdtB protein has cytotoxic DNAse activity (11, 12, 22), suggest that both CdtB and CdtC are likely present in the CDT holotoxin. Our finding that a cdtA mutant still expressed detectable CDT activity does not preclude the presence of CdtA in the CDT holotoxin. The very weak cytotoxic effect detectable in the cdtA mutant would be consistent with CdtA being a structural component essential for fully active toxin. Alternatively, CdtA could function to posttranslationally modify either CdtB or CdtC or both proteins to enable them to function optimally.

This study also confirmed our previous observation that a H. ducreyi cdtC mutant was as virulent as its wild-type parent strain in the temperature-dependent rabbit model (44); this mutant was also recently shown to form pustules in human volunteers at a rate similar to that obtained with its wild-type parent strain (53). In addition, we were unable to demonstrate any decrease in virulence associated with either a cdtA or cdtB mutation (Table 2). RT-PCR analysis of RNA prepared from rabbit lesion material confirmed that all three genes of the H. ducreyi cdtABC cluster were transcribed in rabbit skin during infection. In vivo transcription of at least cdtB was also shown to occur in human volunteers infected with wild-type H. ducreyi (48). These findings suggest that the failure to observe diminished virulence with the H. ducreyi cdtA, cdtB, and cdtC mutants in rabbits and with the cdtC mutant in humans is unlikely to be due to a lack of expression of the relevant gene products in vivo.

The apparent lack of effect of mutations in the cdtABC gene cluster on virulence expression by H. ducreyi in both human and animal models of infection indicates that CDT does not play a role in the development of the pustule that forms in the early stages of chancroid. The determination of whether CDT might be involved in either ulcer development or the retardation of healing of the chancroidal ulcer cannot be accomplished in these model systems. In the former, safety concerns preclude studies involving ulcer development while in the latter model, the lack of ulcer production by the wild-type strain in rabbits will not allow either issue to be addressed. However, data from recent in vitro experiments suggest that CDT could affect the immune response to this pathogen. It has been shown that the spectrum of H. ducreyi CDT activity also includes human T cells (i.e., the Jurkat T-cell line) (14), and this finding is reinforced by a recent study in which H. ducreyi CDT was shown to inhibit proliferation of human T cells and B cells in vitro (47). These results are complemented by studies which indicated that the immunosuppressive factor produced by A. actinomycetemcomitans is a CDT that causes G₂ arrest in human T cells in vitro (42). Therefore, it is conceivable that H. ducreyi CDT might be responsible at some level for the delayed healing of chancroidal ulcers observed in the pre-antibiotic era (38, 50) and for the apparent lack of protective immunity after chancroid (3). Whether H. ducreyi CDT can exert immunosuppressive activity in the human host remains a matter of speculation at this time.