OUP user menu

Variations in lethal toxin and cholesterol-dependent cytolysin production correspond to differences in cytotoxicity among strains of Clostridium sordellii

Daniel E. Voth , Octavio V. Martinez , Jimmy D. Ballard
DOI: http://dx.doi.org/10.1111/j.1574-6968.2006.00287.x 295-302 First published online: 1 June 2006

Abstract

Clostridium sordellii is an emerging human pathogen and frequent contaminant of cadaver-derived tissue transplant material. Herein, we provide data suggesting the potential for severe C. sordellii-associated disease may be dictated by whether the specific strain produces lethal toxin (TcsL) or sordellilysin (SDL), a cholesterol-dependent cytolysin. The virulence factor profiles of 14 C. sordellii isolates were determined, and culture supernatant from six of the isolates was found to be cytotoxic to mammalian cells; yet, only one of these strains conferred cytotoxicity via production of TcsL. Cytotoxicity of TcsL strains correlated with the production of sordellilysin, which was also recognized by an antiperfringolysin O antibody. However, supernatant from TcsL+, SDL strains demonstrated a lower LD50 relative to TcsL, SDL+ strains, suggesting the potential for severe C. sordellii-associated disease may be determined by the particular strain colonizing the host.

Keywords
  • Clostridium sordellii
  • lethal toxin
  • TcsL
  • CDC
  • sordellilysin

Introduction

Clostridium sordellii is a Gram-positive, spore-forming anaerobe, which causes disease in livestock (Lewis & Naylor, 1998; Vatnet al ., 2000a, b) and life-threatening illnesses in humans. Clostridium sordellii has been reported to cause disease in postpartum and postabortion patients and in intravenous drug users (Rorbyeet al ., 2000; Sinaveet al ., 2002; Kimuraet al ., 2004). Additionally, C. sordellii is a common contaminant of cadaver-derived tissue used in transplantation (Malininet al ., 2003; Martinezet al ., 2003) and has been transmitted via such allografts (CDC, 2001).

The severity of C. sordellii-associated disease is due to the organism's ability to grow rapidly and release a variety of soluble virulence factors, which damage host cells. Two large clostridial toxins (LCTs), lethal toxin (TcsL) and hemorrhagic toxin, which glycosylate small GTPases in the cytosol of mammalian cells, are among the primary virulence factors produced by C. sordellii (Boquet, 1999). Unlike other accessory virulence factors produced by C. sordellii, such as hemolysins, neuraminidases, and phospholipases (Arseculeratneet al ., 1969), LCTs are known to demonstrate a remarkably low LD50, ranging from 5 to 50 ng kg−1 in mice (Martinez & Wilkins, 1992), suggesting these toxins would be a major determinant of diseases caused by this pathogen. Furthermore, whether clinical C. sordellii isolates produce one, or combinations, of these virulence factors is currently unknown.

Many pathogenic clostridia are known to produce a plethora of virulence factors, with some species (e.g. strains of Clostridium perfringens) releasing over a dozen different toxins (Hatheway, 1990). In the case of C. perfringens, the specific disease directly correlates with the particular toxins produced by the type infecting the host (Smedleyet al ., 2004). Clinical descriptions of C. sordellii-associated diseases suggest this pathogen acts in a similar manner, as a variety of distinct illnesses have been attributed to this organism (Soper, 1986; Bangsberget al ., 2002; Smedleyet al ., 2004). In some cases, C. sordellii remains localized to the site of infection, with limited myonecrosis, and a shock-like death occurs (Rorbyeet al ., 2000; Sinaveet al ., 2002), while in other reports, bacteremia directly correlates with death of the patient (Moryet al ., 1995; Cunniffe, 1996; Abdulla & Yee, 2000). In cases of postpartum shock, bacteremia can occur, while in other descriptions C. sordellii was not detected in the bloodstream (Bittiet al ., 1997; Rorbyeet al ., 2000). Tissue necrosis also varies among patients with C. sordellii disease, despite similar modes of infection (Browdieet al ., 1975; Soper, 1986; Bangsberget al ., 2002). Thus, different forms C. sordellii disease do not correlate with the mode of infection, suggesting different virotypes may largely contribute to the type of disease occurring in the infected patient. Unfortunately, a collective analysis of the virulence factor profiles of different C. sordellii isolates has not been performed.

For these reasons, we characterized 14 cadaver-derived C. sordellii isolates. The data described herein suggest strain differences are dictated by the presence of TcsL or a previously undescribed hemolysin, sordellilysin (SDL). Collectively, the current findings indicate cadaver-derived isolates of C. sordellii differ substantially in their virulence factor profiles and toxicity, raising the possibility of a range of clinical outcomes in C. sordellii disease depending on the strain present.

Materials and methods

Mammalian cell culture, bacterial strains, and TcsL purification

HeLa, CHO, NIH/3T3, HUV-EC-C, and Raw 264.7 cells were obtained from the American Type Culture Collection (ATCC). HeLa, CHO, and NIH/3T3 cells were maintained in RPMI 1640 medium [10% fetal bovine serum (Gibco)] at 37°C and 6% CO2. HUV-EC-C cells were maintained in Ham's F-12K medium (Gibco) containing 0.1 mg mL−1 heparin (Sigma) and 0.05 mg mL−1 endothelial cell growth supplement (Sigma). Raw 264.7 cells were maintained in Dulbecco's Modified Eagle's Medium (Gibco). All cells were used between passages 10 and 20.

Clostridium sordellii ATCC9714 was obtained from the ATCC, and all other C. sordellii strains were isolated in the microbiology laboratory of the University of Miami Tissue Bank during routine screening of cadaver-derived tissues from clinically nonseptic donors. All samples were collected under aseptic conditions in designated sterile facilities. Isolated organisms were identified using the Microscan Rapid Anaerobic ID Panel (Dade Behring). Isolates were cultured anaerobically at 37°C and TcsL was purified as described previously (Qa'Danet al ., 2001). Unless otherwise indicated, all chemical reagents were obtained from Sigma.

PCR analysis

Oligonucleotides (Integrated DNA Technologies Inc.) (Table 1) were used for PCR amplification of the indicated genes, or gene segments, from each C. sordellii strain. PCR was performed using Expand High Fidelity polymerase (Roche) and appropriate annealing conditions. Each product was cloned into the pGEM/T-Easy vector (Promega), according to the manufacturer's instructions, and analyzed via DNA sequencing. Ribotyping was performed using primers designed to amplify the region between the 16S and 23S rRNA-encoding genes (Cartwrightet al ., 1995) in each isolate. Products were separated via agarose gel electrophoresis, and isolates were assigned to ribotype groups under the criteria of possessing at least one differing band compared with C. sordellii ATCC9714.

View this table:
1

Primers used in this study

PCR reactionPrimers
16S rRNA gene27F=5′-AGAGTTTGATCMTGG-3′
1492R=5′-TACCTTGTTACGACTT-3′
RibotypingForward primer=5′-GCTAACCTTTTAGGAGGCGGC-3′
Reverse primer=5′-GGCTACTTCCTGCACTATTCG-3′
tcslForward primer 1=5′-GACTGACATATGATGAACTTAGTTAACAAAGCCCAA-3′
Reverse primer 1=5′-GACTGAGGATCCTTATACTGTATTTTGAGCAAAATC-3′
Forward primer 2=5′-GACTGACATATGCTTGATAAAGATTATGTTTCTAAA-3′
Reverse primer 2=5′-GACTGAGGATCCTTAGTCTATTTCTGATAATACCAA-3′
Forward primer 3=5′-GACTGACATATGTTTAATAATAATTCAATAACTTTA-3′
Reverse primer 3=5′-GACTGAGGATCCTTACTCACTATTTGCTATAAGAAT-3′
Forward primer 4=5′-GACTGACATATGGAAGATAATCAACGACAAGTTAAA-3′
Reverse primer 4=5′-GACTGAGGATCCTTATTCACTAACTACTAATTCAGC-3′
cdcForward primer=5′-GTACATATCCAGGAGCATTACAAC-3′
Reverse primer=5′-CCACCATTCCCAAGCAAGACCTGT-3′
Southern probe3592F=5′-TTATCTATATATGATGTATTAAATATA-3′
5331R=5′-TATCGCTTTTAAAGACATTAACAA-3′

Cytotoxicity assay

Supernatants (24 and 48 h) were collected from C. sordellii cultures and concentrated using 10 kDa molecular weight cutoff Centricons (Millipore). Supernatant (5 μg well−1) was added in triplicate to a 96-well plate containing 4 × 104 mammalian cells per well. Cells were monitored for 72 h posttreatment for cytopathic effects (CPE) using a 1X-51 Olympus inverted microscope (× 140 magnification). Cell viability was monitored using the CCK-8 Cell Viability Assay (Dojindo Inc.) as described previously (Voth et al., 2004).

Southern blot analysis

A digoxigenin (DIG)-labeled probe was constructed, via PCR, from C. sordellii ATCC9714 DNA using the primers listed in Table 1. The resulting product represented the ∼1.7 kb DNA segment between the NcoI (bp 3592) and PstI (bp 5331) sites on tcsl. This product was labeled using the digoxigenin High Prime DNA Labeling Kit (Roche). Chromosomal DNA was isolated from mid-log phase C. sordellii cultures using the DNeasy Tissue Kit (Qiagen) and restricted overnight at 37°C with NcoI and PstI (New England Biolabs). Following restriction, Southern blotting was performed according to standard procedures.

Immunoblot analysis

Clostridium sordellii ATCC9714 and UMC164 culture supernatants (10 μg) were separated via 10% SDS-PAGE and electro-transferred to a polyvinylidene difluoride membrane (Amersham Pharmacia). The membrane was probed with TcsL or perfringolysin O (PFO) antiserum, and reacting proteins were detected via chemiluminescence (Amersham Pharmacia). TcsL antiserum was isolated following immunization of New Zealand White rabbits with a KLH-conjugated synthetic peptide (TcsL240–260) along with Complete Freund's adjuvant. This 15-residue peptide, IRNLEKFADEDLVRC, represents a putatively exposed hydrophilic region (residues 240–253) at the TcsL N-terminus. Following initial immunization, rabbits were boosted once a month for 3 months with Incomplete Freund's adjuvant plus TcsL240-260 before serum collection. Immunization and serum collection were performed by Sigma Genosys.

Glucosylation and hemolytic assays

TcsL-9714 or TcsL-UMC164 was analyzed in a standard glucosylation assay as previously described (Vothet al ., 2004). Clostridium sordellii supernatants were analyzed in a standard hemolytic assay, using human erythrocytes, as described previously (Tweten, 1988).

LD50 determinations

Eight-week-old female BALB/c mice were injected i.v., via the tail vein, with supernatant from either C. sordellii UMC164 or C. sordellii PS4404. Six mice were injected with each concentration to determine the LD50 value using the method of Reed & Muench (1938).

Results

Characterization of C. sordellii isolates

Each of the isolates examined in this study was confirmed as C. sordellii by 16S rRNA-encoding gene sequence comparisons, and strains were further subgrouped into five ribotypes via amplification of the 16S and 23S rRNA gene intervening sequences (Table 2). Isolates were further examined for the presence of genes encoding TcsL and the production of TcsL, neuraminidase, and phospholipase (Table 2). Only one isolate was similar to the reference strain (ATCC9714) in encoding for, and producing, TcsL (Fig. 1). This toxin, TcsL-UMC164, was purified from dialysis cultures of C. sordellii UMC164 and analyzed via Western blot analysis (Fig. 1d), which further indicated homology to TcsL-9714. Since TcsL is known to modify Ras, Rac, and Cdc42, but not Rho (Voth & Ballard, 2005), TcsL-UMC164 was analyzed for TcsL-specific activity in a standard glucosylation assay. As shown in Fig. 1e, TcsL-UMC164 glucosylated Ras, Rac, and Cdc42, while Rho modification was not observed.

View this table:
2

Phenotypic and genotypic characterization of Clostridium sordellii isolates

Ribotype and strainCharacterization
Ribotype I
 ATCC9714 tcsl +, cdc+, TcsL+, Neu+, SDL, PLC+
Ribotype II
 UMTB2 tcsl , cdc+, TcsL, Neu+, SDL+, PLC+
 UMC178 tcsl , cdc+, TcsL, Neu+, SDL, PLC+
Ribotype III
 UMTB1 tcsl , cdc+, TcsL, Neu+, SDL+, PLC+
Ribotype IV
 UMC164 tcsl +, cdc+, TcsL+, Neu+, SDL, PLC+
 UMC193 tcsl , cdc+, TcsL, Neu+, SDL, PLC+
 UMC212 tcsl , cdc+, TcsL, Neu+, SDL, PLC+
 PS4423 tcsl , cdc+, TcsL, Neu+, SDL, PLC
 PS4475 tcsl , cdc+, TcsL, Neu+, SDL+, PLC+
Ribotype V
 PS4401 tcsl , cdc+, TcsL, Neu+, SDL+, PLC+
 PS4404 tcsl , cdc+, TcsL, Neu+, SDL+, PLC+
 PS4422 tcsl , cdc+, TcsL, Neu+, SDL, PLC+
 PS4451 tcsl , cdc+, TcsL, Neu+, SDL, PLC+
 PS4477 tcsl , cdc+, TcsL, Neu+, SDL, PLC+
 PS4490 tcsl , cdc+, TcsL, Neu+, SDL, PLC+
  • cdc, cholesterol-dependent cytolysin-encoding gene; Neu, neuraminidase; PLC, phospholipase C.

1

Analysis of tcsl and TcsL in Clostridium sordellii isolates. (a) Oligonucleotide primers designed to amplify defined regions of the tcsl gene were included in PCR reactions for each C. sordellii isolate. Amplification products were separated via 0.8% agarose gel electrophoresis and visualized via ethidium bromide staining. Shown are the reactions for ATCC9714 and UMC164. Lane 1=tcsl1−1668, Lane 2=tcsl1669−3468, Lane 3=tcsl3469−5271, and Lane 4=tcsl5272−7095; (b) chromosomal DNA from C. sordellii ATCC9714 and UMC164 was probed for the presence of a ∼1.7 kb region of tcsl via Southern blot analysis. Lanes are designated by the name of the respective C. sordellii strain; (c) purified TcsL from C. sordellii UMC164 was visualized via 10% SDS-PAGE and Coomassie Blue staining. Clostridium sordellii ATCC9714 was visualized for size comparison; (d) purified TcsL from C. sordellii UMC164 was detected via Western blotting using TcsL-specific antiserum. Lanes are designated by the strain name of each C. sordellii isolate; (e) TcsL isolated from C. sordellii UMC164 was included in a standard glucosylation assay with Rho, Ras, Rac, and Cdc42. TcsL from C. sordellii ATCC9714 is included as a control and samples are designated above the gel.

Clostridium sordellii supernatant cytotoxicity

To examine the requirement of TcsL for C. sordellii toxicity, culture supernatant from each C. sordellii isolate was incubated with mammalian cells and observed for cytotoxic activity. Supernatants from seven of the 15 strain supernatants were cytotoxic to HeLa cells; however, cell rounding, indicative of TcsL-like CPE, was only detected upon exposure to supernatants from C. sordellii ATCC9714 and UMC164 (Fig. 2c), even when supernatants from other strains were concentrated 10–100 fold before treatment. The CPE induced by supernatant from strains UMTB1, UMTB2, PS4401, PS4404, and PS4475 differed from TcsL-induced CPE, with cell rounding, blebbing, and lysis occurring within 2 h following treatment (Fig. 2b). In contrast, ATCC9714 and UMC164 supernatants did not induce CPE until ∼24 h following treatment, with cells demonstrating marked rounding without blebbing or detectable lysis (Fig. 2c), suggesting cytolytic strains UMTB1, UMTB2, PS4401, PS4404, and PS4475 induced TcsL-independent cell death.

2

Supernatant cytotoxicity of Clostridium sordellii isolates. Culture supernatant (48 h) from each of the C. sordellii strains was incubated with an ∼75% confluent HeLa cell monolayer in a 96-well plate in a total of 100 μL well−1 and observed for 72 h for cytopathic effects (CPE). Pictures shown are representative of a routine cell treatment with the indicated supernatant, and cells were observed at a magnification of × 140. All hemolytic strains induced similar effects as the PS4404 cell treatment shown here. (a) control cells; (b) PS4404 treatment at 2 h; (c) UMC164 treatment at 24 h.

Characterization of C. sordellii cytolytic activity

The cytolytic activity of TcsL, cytolytic supernatants was similar to that reported for the cholesterol-dependent cytolysins (CDCs) (Tweten, 2005), which induce erythrocyte and tissue culture cell lysis (Tweten, 2005). Thus, TcsL, cytolytic strains were examined in a standard hemolytic assay, which indicated that each of these strains lysed erythrocytes (data not shown). To investigate the nature of this hemolytic factor, these strains were further examined for the presence of a CDC by PCR analysis using primers homologous to pfo, the gene encoding Clostridium perfringens PFO. As shown in Fig. 3a, a ∼1 kb DNA fragment from all 15 C. sordellii strains was amplified using primers homologous to pfo. Supernatant from each isolate was further analyzed using PFO antiserum, which showed that each of the five hemolytic supernatants contained a protein, termed sordellilysin (SDL), which reacted with PFO antiserum (Fig. 3b).

3

Analysis of pfo and sordellilysin in Clostridium sordellii isolates. (a) Oligonucleotide primers designed to amplify cdc based on the reported sequence of pfo were included in PCR reactions for each C. sordellii isolate. Amplification products were separated by 0.8% agarose gel electrophoresis and visualized via ethidium bromide staining. Lanes are designated by the strain names of each isolate in the figure. pRT20, a plasmid containing the pfo-encoding gene, was included as the positive control; (b) 10 μg of each supernatant was separated via 10% SDS-PAGE and proteins were transferred to a polyvinylidene difluoride membrane. Following transfer, the membrane was incubated with a perfringolysin O-specific primary antibody for observation of reacting proteins. Reacting proteins were visualized via enhanced chemiluminescence and the lanes are designated above the blot.

Cell type specificity of TcsL and sordellilysin

Based on initial characterization of the C. sordellii isolates, we reasoned that supernatants from these strains might demonstrate differences in cell tropism and cytotoxicity. Thus, a panel of cell lines, including HeLa, CHO, NIH/3T3, Raw 264.7, and HUV-EC-C cells, was examined for sensitivity to supernatants from the cytotoxic strains. As shown in Fig. 4a, each cell type was sensitive to TcsL, SDL+ supernatant (PS4404) when exposed to 1 and 10 μg of supernatant for 24 h. However, HUV-EC-C cells were substantially more resistant to this cytolytic activity, demonstrating a 40–50% reduction in cell viability at 24 h.

4

Cellular specificity of Clostridium sordellii virulence factors. Culture supernatants from C. sordellii UMC164 (TcsL+, nonhemolytic) and C. sordellii PS4404 (TcsL, hemolytic) were incubated separately with HeLa, CHO, NIH/3T3, Raw 264.7, and HUV-EC-C cells at the indicated concentrations in a total volume of 100 μL well−1. At 24 h posttreatment, cells were examined for cytopathic effects (CPE) by visualization and for viability via CCK-8 staining as described in Materials and methods. Treatments were performed in triplicate and error bars indicate the standard deviation from the mean. (a) Viability of cells incubated with PS4404 supernatant; (b) viability of cells incubated with UMC164 supernatant; (c) CPE of cells incubated with UMC164 supernatant. The % CPE was determined by counting the number of rounded cells in a field of 100 cells in three separate wells for each condition.

As shown in Fig. 4b, each cell type was sensitive to TcsL+, SDL supernatant (UMC164). However, Raw 264.7 cells were only marginally sensitive to this supernatant, with a 15–20% reduction in cell viability at 24 h. This resistance, however, did not correlate with CPE, as Raw 264.7 cells displayed CPE when exposed to 1 and 10 μg of supernatant protein (Fig. 4c). Interestingly, HUV-EC-C cells showed increased sensitivity to supernatant from TcsL+, SDL isolates, with 80–100% CPE occurring at each concentration. Additionally, HUV-EC-C cell viability, in response to TcsL+, SDL supernatant, decreased at each concentration, with an ∼25% decrease when exposed to 0.01 μg of supernatant protein and an ∼65% decrease when exposed to 10 μg of supernatant protein.

LD50 analysis of TcsL+ and SDL+ supernatants

The tissue culture analyses indicated supernatants from the two C. sordellii strains might exhibit differing effects in vivo, as distinct differences exist in tissue tropism for these isolates. To examine the effect of TcsL and sordellilysin on virulence in vivo, C. sordellii supernatant – TcsL+, SDL or TcsL, SDL+– was injected into the tail veins of female BALB/c mice and mice were observed for 24 h. Mice injected with 25–50 μg of TcsL+, SDL supernatant showed lethargy leading to death by 24 h, while mice injected with a 10-fold excess (400–500 μg) of TcsL, SDL+ supernatant demonstrated lethargy leading to death by 24 h (data not shown). Indeed, further analysis revealed an LD50 between 30 and 40 μg for TcsL+ supernatant. These results indicate the importance of TcsL in C. sordellii infection, with TcsL+ supernatant demonstrating much lower LD50 values than SDL+ supernatants.

Discussion

In the current study, virulence factor profiles were determined for 14 C. sordellii isolates. Before this work, C. sordellii has been recognized as a severe pathogen capable of causing disease in a variety of clinical settings; yet, little is known about the phenotypic and genotypic variability of this organism in regard to virulence. The current evidence indicates C. sordellii causes a variety of disease types, suggesting a variation in virulence determinant production. Indeed, several reports suggest C. sordellii is a cause of myonecrotic disease (Soper, 1986; Speraet al ., 1992; Bangsberget al ., 2002), while others associated this organism with a shock-like illness involving limited tissue damage (Bittiet al ., 1997; Rorbyeet al ., 2000; Sinaveet al ., 2002). These differences in C. sordellii-associated disease could be due to the mechanism and site of infection, or to some strains releasing virulence factors that mediate tissue destruction, while other strains produce toxins that induce shock.

In previous reports, C. sordellii isolates have been studied that vary in the production of specific virulence factors, such as TcsH, TcsL, urease, and lecithinase (Nakamuraet al ., 1976, 1983; Popoffet al ., 1985; Greenet al ., 1996). However, the strains examined in these cases were not collected from a single source but, rather, were isolated from various clinical samples. Furthermore, these studies focused on the presence, or absence, of a specific virulence factor in each strain. Thus, in the current study, we sought to characterize the overall production of virulence determinants by C. sordellii strains isolated from a single type of environmental source.

To begin to resolve this issue, we took advantage of the fact that C. sordellii is routinely isolated from cadaver-derived tissue transplant material. Using this single source of C. sordellii isolates, we assessed virulence differences among strains existing in similar host settings. Before virulence factor profile analysis, strains were characterized for basic genotypic and phenotypic differences. Although C. sordellii is considered a highly virulent, toxigenic pathogen, only six of the 14 strains conferred cytotoxic effects in the panel of cell lines tested. While each of the strains produced neuraminidase, and all but one was positive for phospholipase C activity, only the isolates producing TcsL or sordellilysin induced cytotoxic effects. Yet, even the TcsL+ and SDL+ strains demonstrated differences in cell type specificity. Each cell type was sensitive to sordellilysin, which was expected as CDCs are known to lyse most cell types with cholesterol-rich membranes (Tweten, 2005). In contrast, TcsL+ supernatant was cytotoxic but more specific in cell tropism, with endothelial cells (HUV-EC-C) showing enhanced sensitivity to this virulence factor. This is likely due to either limitations in cell surface receptor number or the dependence of different cells on small GTPase-targeted signaling pathways. The extensive sensitivity of endothelial cells to TcsL+ supernatants is of particular interest, as C. sordellii causes shock-like illness in postpartum women and sudden death in livestock (McGregoret al ., 1989; Lewis & Naylor, 1996; Bittiet al ., 1997; Rorbyeet al ., 2000). In future studies, it will be important to characterize isolates involved in shock from C. sordellii infections to determine if these are TcsL+ strains.

The role of sordellilysin in virulence is under current investigation. Based on our LD50 analyses, sordellilysin demonstrated a ∼10-fold higher LD50 in mice compared with TcsL. The absence of a pronounced in vivo effect of sordellilysin is not unusual given what is known regarding CDCs. Although these toxins are produced by numerous Gram-positive pathogens, their overall contribution to disease remains unclear. Indeed, in the case of PFO, mutants lacking this CDC show only minor alterations in virulence (Awadet al ., 2001). It is worth noting that a pfo-like gene was detected in all strains examined in this study; however, the toxin was only expressed in five of these isolates, suggesting that many strains may have attenuated the expression of sordellilysin as the organism resides as part of the normal bowel flora.

While the current study focused on characterizing the virulence profile of C. sordellii in the context of tissue transplant material, these findings can be extrapolated to other types of disease caused by this organism. With the exception of intravenous drug users, C. sordellii disease is initiated from endogenous sources, likely through the relocalization of bowel flora to the site of tissue damage. Thus, these strain variations suggest differences that could contribute to specific types of C. sordellii disease. This may also explain why C. sordellii disease is sporadic, yet, invariably severe. One would predict that postpartum contamination of damaged tissue with bowel flora is not particularly uncommon; yet, C. sordellii causes disease infrequently in this clinical setting (McGregoret al ., 1989; Bittiet al ., 1997; Rorbyeet al ., 2000; Sinaveet al ., 2002). Furthermore, a shock-like disease caused by C. sordellii has been reported in three women taking RU486 (CDC, 2005). While still speculative, based on our findings, these cases may involve a combination of postpartum or postabortion tissue damage and women colonized with a virulent strain of C. sordellii. For these reasons, it will be of particular interest to examine isolates of C. sordellii for their virulence profiles and correlate this with disease severity.

Acknowledgements

This work was supported by The State of Oklahoma and NIH NCRR grant RR15564 to J.D.B. We would like to acknowledge Dr T.I. Malinin and Dr B.E. Buck (University of Miami) for the collection of specimens for study, Dr Rodney Tweten (University of Oklahoma Health Sciences Center) for the generous gift of pRT20 and PFO antiserum, and Dr Marvin Whiteley for assistance with Southern blotting.

References

View Abstract