Report

A Bacterial Toxin That Controls Cell Cycle Progression as a Deoxyribonuclease I-Like Protein

See allHide authors and affiliations

Science  13 Oct 2000:
Vol. 290, Issue 5490, pp. 354-357
DOI: 10.1126/science.290.5490.354

Abstract

Many bacterial pathogens encode a multisubunit toxin, termed cytolethal distending toxin (CDT), that induces cell cycle arrest, cytoplasm distention, and, eventually, chromatin fragmentation and cell death. In one such pathogen, Campylobacter jejuni, one of the subunits of this toxin, CdtB, was shown to exhibit features of type I deoxyribonucleases. Transient expression of this subunit in cultured cells caused marked chromatin disruption. Microinjection of low amounts of CdtB induced cytoplasmic distention and cell cycle arrest. CdtB mutants with substitutions in residues equivalent to those required for catalysis or magnesium binding in type I deoxyribonucleases did not cause chromatin disruption. CDT holotoxin containing these mutant forms of CdtB did not induce morphological changes or cell cycle arrest.

Campylobacter jejuni, the most common cause of food-borne infectious illnesses in the United States (1), encodes a toxin termed CDT, which is considered to be an important virulence factor (2–4). This toxin causes eukaryotic cells to arrest in the G2/M transition phase of the cell cycle (5). Intoxicated cells show a characteristic accumulation of the phosphorylated form of the cell cycle regulator CDC2, as well as an increase in their DNA content (4N), consistent with a cell cycle blockage at the G2/M boundary (6–10). After intoxication, cell division stops, but the cytoplasm continues to grow and distend, resulting in cells up to five times their normal size. Intoxicated cells maintain viability for extended periods of time, although they eventually show morphological changes in the chromatin such as condensation and/or fragmentation and ultimately die. In addition toC. jejuni, CDT-homologous toxins have been described in several other important bacterial pathogens (11–17), but little is known about their mechanism of action. CDTs are encoded by a cluster of three highly conserved genes of unknown function, cdtA, cdtB, andcdtC (4).

To investigate which of the Cdt subunits may have toxic activity within the host cell, we transfected COS-1 cells with plasmids expressing epitope-tagged CdtA, CdtB, or CdtC or vector control (18). Cells expressing CdtA or CdtC or transfected with the vector control displayed an apparently normal morphology and showed no signs of intoxication (Fig. 1). In contrast, cells transfected with a plasmid expressing CdtB exhibited striking alterations characterized by fragmented nuclei and often a total collapse of the chromatin (Fig. 1). Changes in the chromatin were apparent as early as 18 to 24 hours after transfection when nuclei of transfected cells began to exhibit a distinct smooth appearance (Fig. 1). Forty-eight hours after transfection, the nuclei of transfected cells appeared seriously compromised, exhibiting marked fragmentation or outright disappearance of the chromatin (Fig. 1). Thus, CdtB was responsible, at least in part, for the activity of the CDT toxin within the host cell.

Figure 1

Effect of the transient expression of CDT toxin components in cultured cells. COS-1 cells were transfected with vectors coding for M45 epitope– tagged C. jejuni CdtA, CdtB, or CdtC toxin subunits. Twenty-four or 48 hours after transfection, cells were stained with a monoclonal antibody directed to the M45 epitope tag to visualize cells expressing the individual toxin subunits and with DAPI to visualize the chromatin (18). Images were obtained with a Nikon Diaphot inverted fluorescence microscope fitted with a Princeton Instruments Micromax digital camera. Scale bar, 50 μm.

We examined the predicted amino acid sequence of this toxin subunit in an effort to identify clues about its function. Amino acid sequence comparison revealed similarity to deoxyribonuclease (DNase) I–like proteins. Further alignment of CdtB with members of the DNase I protein family revealed a striking conservation of most residues that mutagenesis and structural analysis have shown to be essential for enzymatic activity (19–21) (Fig. 2A). These include residues that are part of the active site as well as residues that are important for Mg2+ binding, an essential requirement for the catalytic activity of this family of proteins (21). Further sequence comparison revealed that these residues are absolutely conserved in all CdtB proteins from other bacterial pathogens.

Figure 2

(A) Amino acid sequence comparison of C. jejuni CdtB toxin subunit with members of the DNase-I family of related proteins (33). Residues that have been shown to be essential for catalysis are boxed. Arrows indicate the CdtB residues that were mutated for these studies. The alignment was done with the ClustalW program (34). The DHP-1, DHP-2, DHP-3, DNase I, and DNase X sequences correspond to human DNases. (B) Effect of the transient expression of CdtB mutants in cultured cells. COS-1 cells were transfected with vectors coding for different M45 epitope–tagged CdtB mutants as indicated (18). Forty-eight hours after transfection, cells were stained with a monoclonal antibody directed to the M45 epitope tag to visualize cells expressing the different CdtB proteins and with DAPI to visualize the chromatin. Images were obtained with a Nikon Diaphot inverted fluorescence microscope fitted with a Princeton Instruments Micromax digital camera. Scale bar, 50 μm.

To confirm the putative DNase activity of CdtB, we constructed mutant derivatives of this toxin subunit carrying single amino acid substitutions in residues that have been shown to be critical for the catalytic activity of members of the type I DNase family of proteins (22). The residues that were mutated were equivalent to those shown in other type I DNases to be either components of the active site (i.e., His152) or essential for magnesium binding (i.e., Asp185) (19–21) (Fig. 2A). Plasmids encoding either epitope-tagged CdtBH152Q or CdtBD185S were transfected into COS-1 cells, and 48 hours after transfection, the cells were stained with an antibody directed to the epitope tag and with 4′,6′-diamidino-2-phenylindole (DAPI) to examine the structure of the chromatin. Cells transfected with either of these plasmids exhibited completely normal nuclear morphology despite the presence of high levels of mutant CdtB protein as judged by the bright fluorescence staining and Western blot analysis (Fig. 2B). This was in sharp contrast to cells transfected with wild-type CdtB, which showed a complete collapse of their chromatin (Fig. 2B). The mutant CdtB proteins exhibited exclusive nuclear localization, implicating the nucleus as the site of action for this toxin subunit and consistent with its potential role as a DNase (Fig. 2B). Because CdtB does not exhibit a detectable consensus nuclear localization signal (NLS), its mechanism of translocation to the nucleus may involve either an atypical NLS or a carrier protein (23).

To confirm the toxic activity of CdtB, we purified this protein to homogeneity and microinjected it into cells (24) (Fig. 3B). Microinjection into COS-1 or REF52 cells of a 1 mg/ml solution of purified CdtB resulted in marked changes in the chromatin as early as 4 hours after microinjection (Fig. 3A). In contrast, microinjection of equal amounts of CdtBH152Q purified in identical fashion did not cause any visible alteration to the chromatin of microinjected cells (Fig. 3A). Both wild-type CdtB and CdtBH152Q were exclusively localized to the nucleus (Fig. 3A).

Figure 3

Effect of the microinjection of purified CdtB or CdtBH152Q in cultured cells. (A) COS-1 or REF52 cells grown on gridded cover slips were microinjected with a 1 mg/ml solution of purified preparations of CdtB or CdtBH152Q (24). Four hours after microinjection, cells were fixed and stained with an antibody directed to CdtB and with DAPI to visualize the chromatin. In some experiments, microinjected cells were identified by comicroinjection of FITC-labeled dextran (24). (B) COS-1 or REF52 cells grown on gridded cover slips were microinjected with a 1 μg/ml solution of purified preparations of CdtB (25). Four days after microinjection, cells were fixed and stained with DAPI to visualize the chromatin. Microinjected cells were identified by comicroinjection of FITC-labeled dextran (25). Notice the marked enlargement of the nuclei and the distention of the cytoplasm in microinjected cells. Similar results were obtained in several repetitions of this experiment (25). Images were obtained with a Nikon Diaphot inverted fluorescence microscope fitted with a Princeton Instruments Micromax digital camera. Scale bar, 50 μm. A coomassie blue–stained SDS-polyacrylamide gel of the purified CdtB protein preparations used in the microinjection studies is shown.

The morphological changes induced by the overexpression of CdtB or the microinjection of a 1 mg/ml solution of purified CdtB do not completely resemble those changes induced by CDT holotoxin intoxication. We hypothesized that this difference could be the result of the vastly different intracellular levels of CdtB under these two different conditions. To test this hypothesis, we microinjected into COS-1 and REF52 cells increasingly lower amounts of purified CdtB protein and examined the effects on cellular and nuclear morphology (25). Microinjection of a solution containing purified CdtB (1 μg/ml) resulted in vast distention of the cell cytoplasm and severe enlargement of the nucleus 4 to 5 days after microinjection (Fig. 3B). These changes closely resemble those induced by the CDT holotoxin, indicating that, when microinjected into cells, CdtB by itself was capable of recapitulating the toxic effects observed by CDT holotoxin treatment. As expected, microinjection of similar amounts of CdtBH152Q did not result in any detectable cellular changes (25).

To investigate the importance of the DNase activity of CdtB on the toxicity of the CDT holotoxin, we constructed plasmids that encode the mutant proteins CdtBH152Q or CdtBD185S along with CdtA and CdtC in the same genetic organization as that of the wild-type locus (26). Extracts fromEscherichia coli strains carrying these plasmids or the wild-type cdtABC locus were prepared, and their toxicity was examined by a variety of assays (27). The levels of the CdtB mutant proteins as well as those of the other components of the holotoxin were equivalent in all the extracts used in these studies (Fig. 4C). Extracts from E. coli expressing wild-type CdtB induced marked morphological changes in intoxicated cells characterized by a marked distention and enlargement of the cell cytoplasm and nucleus (Fig. 4A). In contrast, cells treated with extracts from E. coli expressing the catalytic (CdtBH152Q) or the Mg2+-binding (CdtBD185S) mutants exhibited the same morphology as cells treated with extracts from E. coli carrying the vector alone (Fig. 4A). To examine the effect of the mutant toxins on cell cycle progression, we analyzed by flow cytometry the DNA content of cells treated with the same extracts of E. coli expressing wild-type or mutant forms of CdtB. Consistent with a G2/M block, cells intoxicated with extracts from bacteria expressing wild-type holotoxin exhibited a 4N DNA content (Fig. 4B). In contrast, the flow cytometric profiles of cells infected with extracts from bacteria expressing either of the mutant forms of CdtB (CdtBH152Q or CdtBD185S) were indistinguishable from the profile of untreated cells or cells treated with extracts from bacteria carrying the empty vector (Fig. 4B). Thus, CdtB DNase activity is essential for CDT toxicity.

Figure 4

Effect of mutations in the predicted catalytic and Mg2+-binding sites of CdtB on CDT toxicity. (A) Morphology of Henle-407 intestinal epithelial cells after treatment with CDT holotoxin containing either mutant or wild-type CdtB. Henle-407 cells were treated with different extracts ofE. coli expressing the C. jejuni CDT toxin containing either wild-type CdtB or its mutant derivatives CdtBH152Q or CdtBD185S. As a control, cells were treated with an extract of E. coli carrying the empty vector (vector). Images were captured 48 hours after treatment with a Nikon Diaphot inverted microscope fitted with a Princeton Instruments Micromax digital camera. Scale bar, 50 μm. (B) Cell cycle progression of Henle-407 intestinal epithelial cells after treatment with CDT toxin containing different CdtB mutants. Henle-407 cells were treated in an identical manner and with the same preparations described in (A). Seventy-two hours after treatment, cells were fixed, stained, and examined for DNA content by flow cytometry (27). The peaks corresponding to cells in G0-G1, S, or G2 are indicated. A minimum of 10,000 nuclei per sample were analyzed. (C) Expression of CdtA, CdtC, and wild-type CdtB or its mutant derivatives CdtBH152Q or CdtBD185S in the extracts used in experiments described in (A) and (B). Whole-cell lysates (labeled A) and sonicated extracts (labeled B) were separated by SDS-polyacrylamide gel electrophoresis, and the levels of the CdtA, CdtB, and CdtC proteins were evaluated by Western immunoblot with specific antibodies.

DNA damage triggers a series of carefully controlled processes that stop cell cycle progression to ensure that cell division will not proceed to the next phase until the DNA damage has been repaired (28). The G2/M cell cycle arrest induced by CDT toxins is therefore most likely the consequence of DNA damage inflicted by the DNase activity of its CdtB subunit upon delivery into the cell. Although DNA damage can lead to cell cycle arrest in either G1/S or G2/M, CDT-intoxicated cells invariably arrest in G2/M. A plausible explanation for this observation may be that CdtB DNase activity is directed to single-stranded DNA only present during the S phase of the cell cycle, which immediately precedes G2. Alternatively, CdtB may have access to the chromatin only in G2 or may require a cofactor for its activity that is only available in this phase of the cell cycle. Consistent with either hypothesis, purified CdtB exhibited only very poor DNase activity in vitro, which was only detectable on single-stranded DNA templates.

Bacterial infections are increasingly considered a potential predisposing factor for the development of cancer. The presence of a bacterial toxin capable of causing DNA damage in a commonly occurring intestinal pathogen such as C. jejuni may not only aid its pathogenicity but may constitute a predisposing factor for intestinal cancer.

  • * To whom correspondence should be addressed. E-mail: jorge.galan{at}yale.edu

REFERENCES AND NOTES

View Abstract

Navigate This Article