We have previously shown that Actinobacillus actinomycetemcomitans produces an immunosuppressive factor that is encoded by the cdtB gene, which is homologous to a family of cytolethal distending toxins (Cdt) expressed by several Gram-negative bacteria. In this study, we report that the cdt locus in A. actinomycetemcomitans is composed of five open reading frames, designated orf1, orf2, cdtA, cdtB, and cdtC. The deduced amino acid sequences of the five open reading frames are highly conserved among A. actinomycetemcomitans strains 652, Y4, 29522, and HK1651. There is also strong homology with the Cdt proteins of Haemophilus ducreyi (87–91%), but only partial homology with that of Campylobacter jejuni and Escherichia coli (29–48%). Analysis of A. actinomycetemcomitans mRNA by RT-PCR suggests that the two small open reading frames upstream of cdtA are coexpressed with cdtA, cdtB, and cdtC. We next utilized a series of plasmids that express various combinations of the cdt genes to determine their requirement for expression of immunoinhibitory activity. Cell extracts of E. coli transformed with each of the plasmids were tested for their capacity to induce G2 arrest in the cell cycle of PHA-activated human T cells. These experiments suggest that expression of cdtB alone is sufficient to induce G2 arrest in human T cells, but do not exclude the possibility that cdtC also contributes to cell cycle arrest. The implications of our results with respect to the function of the individual Cdt proteins are discussed.

SummaryWe have previously shown that Actinobacillus actinomycetemcomitans cytolethal-distending toxin (Cdt) is a potent immunosuppressive agent that induces G2/ M arrest in human lymphocytes. In this study, we explored the possibility that Cdt-mediated immunotoxicity involves lipid membrane microdomains. We first determined that following treatment of Jurkat cells with Cdt holotoxin all three Cdt subunits localize to these microdomains. Laser confocal microscopy was employed to colocalize the subunits with GM1-enriched membrane regions which are characteristic of membrane rafts. Western blot analysis of isolated lipid rafts also demonstrated the presence of Cdt peptides. Cholesterol depletion, using methyl b -cyclodextrin, protected cells from the ability of the Cdt holotoxin to induce G2 arrest. Moreover, cholesterol depletion reduced the ability of the toxin to associate with Jurkat cells. Thus, lipid raft integrity is vital to the action of Cdt on host cells. The implications of our observations with respect to Cdt mode of action are discussed.

We have shown the Actinobacillus actinomycetemcomitans produces an immunosuppressive factor encoded by the cytolethal distending toxin (cdt)B gene, which is homologous to a family of Cdts expressed by several Gram-negative bacteria. We now report that the capacity for CdtB to induce G2 arrest in Jurkat cells is greater in the presence of the other Cdt peptides: CdtA and CdtC. Plasmids containing the cdt operon were constructed and expressed in Escherichia coli; each plasmid contained a modified cdt gene that expressed a Cdt peptide containing a C-terminal His tag. All three Cdt peptides copurified with the His-tagged Cdt peptide. Each of the peptides associated with the complex was truncated; N-terminal amino acid analysis of CdtB and CdtC indicated that the truncation corresponds to cleavage of a previously described signal sequence. CdtA was present in two forms in crude extracts, 25 and 18 kDa; only the 18-kDa fragment copurified with the Cdt complexes. Cdt complexes were also immunoprecipitated from A. actinomycetemcomitans extracts using anti-CdtC mAb. Exposure of Jurkat cells to 40 pg resulted in >50% accumulation of G2 cells. CdtB and CdtC were detected by immunofluorescence on the cell surface after 2-h exposure to the holotoxin. CdtA was not detected by immunofluorescence, but all three peptides were associated with Jurkat cells when analyzed by Western blot. These studies suggest that the active Cdt holotoxin is a heterotrimer composed of truncated CdtA, CdtB, and CdtC, and all three peptides appear to associate with lymphocytes.

We have previously shown that Actinobacillus actinomycetemcomitans produces an immunosuppressive factor encoded by the cytolethal distending toxin (cdt)B gene. In this study, we used rCdt peptides to study the contribution of each subunit to toxin activity. As previously reported, CdtB is the only Cdt subunit that is capable of inducing cell cycle arrest by itself. Although CdtA and CdtC do not exhibit activity alone, each subunit is able to significantly enhance the ability of CdtB to induce G2 arrest in Jurkat cells; these effects were dependent upon protein concentration. Moreover, the combined addition of both CdtA and CdtC increased the ED50 for CdtB >7000-fold. In another series of experiments, we demonstrate that the three Cdt peptides are able to form a functional toxin unit on the cell surface. However, these interactions first require that a complex forms between the CdtA and CdtC subunits, indicating that these peptides are required for interaction between the cell and the holotoxin. This conclusion is further supported by experiments in which both Jurkat cells and normal human lymphocytes were protected from Cdt holotoxin-induced G2 arrest by pre-exposure to CdtA and CdtC. Finally, we have used optical biosensor technology to show that CdtA and CdtC have a strong affinity for one another (10−7 M). Furthermore, although CdtB is unable to bind to either CdtA or CdtC alone, it is capable of forming a stable complex with CdtA/CdtC. The implications of our results with respect to the function and structure of the Cdt holotoxin are discussed.

A geographically homogeneous population of 83 subjects, from 21 families with localized juvenile periodontitis (IJP), and 35 healthy control subjects was monitored, over a 5-year period, for the presence of the periodontal pathogen ActinobaciUlus actinomycetemcomitans. Restriction fragment length polymorphism (RFLP) analysis was used to monitor the distribution of genetic variants of this bacterium in UJP-susceptible subjects that converted from a healthy to a diseased periodontal status. A. actinomycetemcomitans was cultured from 57% of the UP family members accessioned into the study. Nine of 36 LJP-susceptible subjects, in seven families, developed signs of periodontal destruction. All but one of these conversion subjects harbored A. actinomycetemcomitans. Bacterial variants representative of a single RFLP group (II) showed the strongest correlation with conversion (P < 0.002). Six of nine conversion subjects were infected with A. actinomycetemcomitans from this group. RFLP group II variants also prevailed in 8 of 22 probands but were absent in the 35 healthy control subjects. In contrast to the selective distribution of group II variants in diseased individuals, variants belonging to RFLP groups XIII and XIV were found exclusively in the control subjects. Thus, the use of RFLP to type clinical isolates of A. actinomycetemcomitans has resulted in the identification of genetic variants that predominate in UIP and health. These results indicate that studies concerned with the pathogenicity of this bacterium in IJP should be focused on the group II variants.