| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1 Agricultural Institute of Slovenia, Hacquetova 17, Ljubljana 1000, Slovenia
2 Université Catholique de Louvain, Faculté de Médicine, Unité de Microbiologie, Avenue Hippocrate 54 90, B-1200 Brussels, Belgium
3 Institute of Public Health Maribor and University of Maribor, Medical Faculty, Slomskov trg 15, 2000 Maribor, Slovenia
Correspondence
Maja Rupnik
maja.rupnik{at}uni-mb.si
Received 6 September 2006
Accepted 6 November 2006
Abbreviations: LCT, large clostridial toxin; PaLoc, pathogenicity locus; TcdA, toxin A; TcdB, toxin B.
 |
INTRODUCTION |
|---|
 TOP INTRODUCTION METHODSRESULTS AND DISCUSSIONREFERENCES |
|---|
Variability in the tcdA and tcdB genes has been well studied and the strains with changed genes have been grouped into toxinotypes (Rupnik et al., 1998). Major changes are defined as deletions and RFLPs in all six PCR fragments covering tcdA and tcdB, whilst minor changes are limited to only one of the PCR fragments (usually the A3 fragment) covering the toxin genes or other pathogenicity locus (PaLoc) regions. To date, 22 toxinotypes have been described and two additional types are reported in this paper. The best-studied and the most widespread are toxinotypes III (including ribotypes 027, 034, 075 and 080) and VIII (including ribotypes 017 and 047) (Rupnik et al., 2001).
Binary toxin CDT is detected in some strains in addition to TcdA and TcdB, the primary virulence factors of C. difficile (Popoff et al., 1988; Rupnik et al., 2003a). Binary toxin CDT is composed of two unlinked components, the enzymic (CDTa) and binding (CDTb) components, and acts as an actin-specific ADP-ribosyltransferase (Popoff et al., 1988; Barth et al., 2004). CDT has been shown to have a cytotoxic effect on Vero cells in culture (Perelle et al., 1997) and an enterotoxic effect in a rabbit ileal loop assay (Geric et al., 2006). The role of CDT in the development of human disease is not currently well understood. Interestingly, binary toxin-producing strains often seem to be associated with community-acquired C. difficile infections (Terhes et al., 2004; Barbut et al., 2005). The prevalence of binary toxin-producing strains has been estimated to be between 1.6 and 6 % (Stubbs et al., 2000; Rupnik et al., 2001, 2003b; Geric et al., 2003; Goncalves et al., 2004; Terhes et al., 2004; Alonso et al., 2005), although much higher percentages have been reported in other studies (Braun et al., 2000; Perelle et al., 1997; Spigaglia & Mastrantonio, 2002; McEllistrem et al., 2005) and the incidence seems to have increased in recent years (Spigaglia & Mastrantonio, 2004).
CDT is encoded on a 4.3 kb chromosomal locus, composed of the genes cdtA and cdtB (Perelle et al., 1997). Within the fully sequenced genome of A+B+CDT– C. difficile strain 630 (Sebaihia et al., 2006), regions with significant similarity to binary toxin have been found to be present, representing a truncated CDT locus (Fig. 1
). The regions at the 5' end of the cdtA gene and the 3' end of the cdtB gene were highly conserved, but approximately 2 kb of sequence was missing in this strain (Fig. 1a) (Rupnik et al., 2003a). A similar 2 kb deletion has been reported for the reference C. difficile strain VPI 10463 (Chang & Song, 2001).
|
The aim of this study was to determine the distribution of the truncated and full-length CDT locus in C. difficile strains with different large clostridial toxin (LCT) production status and from different toxinotypes and to assess the correlation of the truncated tcdC gene with different variant toxinotypes.
|
METHODS |
|---|
|
TOPINTRODUCTIONMETHODSRESULTS AND DISCUSSIONREFERENCES |
|---|
Eighty strains were from toxinotype 0 (A+B+ strains producing toxin TcdA and TcdB, identical to the reference strain VPI 10463), whilst 25 did not produce TcdA and TcdB (A–B–). The majority of these strains were from the collection of M. Delmée at UCL Brussels and belonged to different serogroups (A–I, K and X). Additional strains from various origins (Sweden, Germany and Japan) previously sent to our laboratory for toxinotyping were also included and were subsequently serotyped as described by Delmée et al. (1985).
The remaining 41 strains were representatives of all 24 known toxinotypes (http://www.mf.uni-mb.si/mikro/tox; 22 toxinotypes published so far, with two new ones described in this paper). For each toxinotype, the reference strain was tested. For toxinotypes IV, V, VI, IX, XIa, XIb and XII we tested one additional strain, for toxinotype III two additional strains and for toxinotype VIII seven additional strains. Three strains of toxinotype III were representatives of subtypes IIIa, IIIb and IIIc, which differed in B2s fragment RFLPs (Geric et al., 2004). Two representatives of the so far unpublished toxinotypes determined in this study (XXIII and XXIV) were also included. Strain 8785 originated from the collection of M. Delmée, Brussels, Belgium, and was the type strain of serogroup A9. It had a novel combination of restriction patterns in fragments B1 (restriction pattern 5) and A3 (restriction pattern 2) and therefore represents a novel toxinotype, XXIII. Strain 597B, originally isolated in Kuwait, is the type strain of toxinotype XXIV. It showed no differences in the tcdB and tcdA genes when compared with C. difficile reference strain VPI 10463, but was distinguished from toxinotype 0 by a small deletion (
20 bp) in the tcdC gene and by the presence of the binary toxin genes cdtA and cdtB (see Results and Discussion).
Characterization of truncated types of tcdC. The tcdC gene was amplified by PCR for the PLC fragment using primers Tim2 and Struppi2, as described previously (Braun et al., 1996), and was digested with NcoI to reveal short deletions within the tcdC gene.
CDT locus amplification. Strains were grown on Columbia 5 % sheep blood agar (BBL; BD). For crude DNA preparation, 12–24 h cultures were suspended in 5 % Chelex 100 (Bio-Rad), vortexed and boiled for 10 min. After centrifugation for 10 min at 10 000 g, the supernatant was transferred into a fresh tube and stored at 4 °C until use.
PCR for the BTb fragment corresponding to a shorter part of the cdtB gene encoding the binding component of the binary toxin (Fig. 1a
) was performed as described previously (Stubbs et al., 2000).
Primers BIN5 (5'-AATATTGGGAGGGAGAATAAATG-3') and BIN8 (5'-AATATATATTGTATTGAGGGGAC-3'), which are positioned outside the CDT locus (Spigaglia & Mastrantonio, 2002), were used for detection of the complete CDT locus sequence. PCRs were performed in a total volume of 50 µl using 3 µl crude DNA extract, 200 µM each dNTP and 15 pM each primer. The final concentration of MgCl2 in the reaction mixture was 3 mM. The hot-start PCR protocol started with denaturation at 93 °C for 3 min after which 2 U Taq polymerase (Roche) was added per reaction, followed by 35 cycles of denaturation at 93 °C for 4 s and annealing and extension at 47 °C for 5 min. The reaction was terminated by a 10 min incubation at 47 °C. Fragments were visualized on a 1 % agarose gel.
Similarity dendrogram of the PaLoc regions. A dendrogram of the PaLoc region for the different toxinotypes was constructed based on PCR RFLP patterns in the reference strains of toxinotypes 0–XXIV. For each toxinotype, data on amplification and RFLPs for seven DNA fragments covering the PaLoc region (Braun et al., 1996; Rupnik et al., 1998; Geric et al., 2004) were included in the matrix: B1 cut with HincII and AccI, B2s cut with HindIII and RsaI, B3s cut with HindIII and RsaI, A1 cut with NsiI and PstI, A2 cut with HaeIII, A3 cut with EcoRI and PLC cut with NcoI.
For successful amplification of a particular fragment as well as the presence of each possible restriction fragment, ‘1’ was entered in the matrix and for its absence ‘0’ was recorded. Jaccard's similarity coefficients were calculated using the computer software Free-Tree version 0.9.1.50 (Pavlycek et al., 1999) and analysed by the unweighted pair group method using arithmetic averages (UPGMA). The resulting PaLoc similarity tree of discrete characters was created using MEGA3 (Kumar et al., 2004).
|
RESULTS AND DISCUSSION |
|---|
|
TOPINTRODUCTIONMETHODSRESULTS AND DISCUSSIONREFERENCES |
|---|
) used for routine detection of functional CDT genes, one strain (597B) was positive and subsequently was found to contain the full-length CDT locus. The same strain also had an altered tcdC gene (Fig. 2; type 2). As toxinotypes are defined as groups of strains with changes in the PaLoc region (and not solely changes in the tcdA and tcdB genes), we included this strain as a novel toxinotype, XXIV. Characteristics of this toxinotype were a deletion in the tcdC gene and the presence of binary toxin genes, with the tcdA and tcdB genes being identical to toxinotype 0. To date, only one other similar strain has been described that was typed as toxinotype 0 using A3 and B1 fragment PCR, but had the binary toxin genes and a tcdC deletion (Spigaglia & Mastrantonio, 2002).
|
Forty-one strains from variant toxinotypes were tested for deletions in tcdC. After restriction with NcoI, four restriction types could be distinguished: type 1 was the full-length gene and types 2–4 represented three different deletions (Fig. 2). So far in the literature, tcdC deletions have been described in toxinotypes V and VI (Spigaglia & Mastrantonio, 2002) and toxinotype III (Warny et al., 2005; McDonald et al., 2005; Loo et al., 2005). Some toxinotype III strains have an additional nonsense mutation in the tcdC gene (MacCannell et al., 2006). As shown in Fig. 3, the tcdC deletions were not specific for toxinotypes III, V and VI, but were present in most of the variant toxinotypes. Truncated versions 2 and 3 were prevalent, whilst version 4 was detected only in toxinotype IV. In two toxinotypes with large deletions in the tcdA gene (X and XVII), tcdC could not be amplified with the primers used.
|
). The short 2.3 kb version of the CDT locus has been reported previously in strains VPI 10463 and 630 (Chang & Song, 2001; Rupnik et al., 2003a) and does not lead to expression of functional binary toxin. Our data suggested that the deletion in the truncated CDT locus is always of the same size (Fig. 1b). Some minor sequence differences in CDT genes have been published (Stubbs et al., 2000; Pituch et al., 2005), but whether they affect the function of the toxin is not known. The truncated versions have not been described so far for other species known to produce related binary toxins (Clostridium perfringens type E and Clostridium spiroforme).
It has been recognized previously that the presence of functional binary toxin genes encoded by the 4.3 kb CDT locus correlates well with variant toxinotypes including A+B+, A–B+ and A–B– strains (Stubbs et al., 2000; Rupnik et al., 2001, 2003a, b; Geric et al., 2004; Spigaglia & Mastrantonio, 2002; Goncalves et al., 2004). This study presents an overview of all of the known variant toxinotypes and has confirmed that the full-length CDT locus is found in variant toxinotypes with significantly changed LCT toxin genes (shown in bold in Table 1), whereas the truncated version of the CDT locus is predominant in variant toxinotypes with minor changes in the LCT toxin genes (Table 1). Only three toxinotypes did not contain any CDT-related sequences using the primers and amplification conditions described: II, VIII and XXI. Interestingly, two of them (VIII and XXI) are major toxinotypes. For type XXI, only a single representative is known. However, toxinotype VIII represents the most widespread variant toxinotype (A–B+ strains of serogroups F and X); therefore, a larger selection of strains could be tested. However, all were found to be negative for CDT amplification.
|
). This tree did not show the phylogenetic relationships among toxinotypes, but clearly distinguished among toxinotypes with minor changes (branch 1) and major changes (branches 2 and 3) in the PaLoc region. Toxinotypes with similar changes in the PaLoc region grouped together, e.g. V–VII, XVI and XI, and X and XVII. Interestingly, toxinotypes XXI and VIII, the only major variant toxinotypes without the CDT locus, were positioned separately and did not belong to any of the major branches. The evolution of the C. difficile CDT locus needs further investigation, especially with regard to the striking co-existence of binary toxin genes and changes in the PaLoc region in variant strains, and it has been suggested that different toxinotypes do not have a common ancestor (Wozniak et al., 2000).
In non-variant strains (toxinotype 0), the truncated CDT locus was found in 63/80 strains (79 %) (Table 1). In several cases where more than one strain per serogroup was tested (e.g. serogroups K and A14), there was no correlation between serogroup and the type of CDT locus (data not shown). Neither the full-length nor the truncated form of the CDT locus was amplified in any of the 25 PaLoc-negative A–B– strains (Table 1). However, A–B–CDT+ strains have been described previously and two subgroups have been differentiated. The first group is A–B–CDT+ strains in which the PaLoc is replaced by 115 bp, as has been found for non-toxinogenic strains (Geric et al., 2003). The second group of A–B–CDT+ strains is represented by toxinotypes XIa and XIb, known to produce binary toxin but not toxins TcdA and TcdB, although a short part of the non-functional PaLoc is still present (Table 1) (Stubbs et al., 2000; Geric et al., 2003).
The presence of the full-length or truncated CDT locus in A+B+ strains and its predominant absence in A–B+ and A–B– strains could suggest a possible positive selection of strains with numerous virulence factors.
In summary, we have presented data showing the heterogeneity of the tcdC gene and the binary toxin locus in C. difficile strains. Both features – deletions in tcdC and the presence of a functional full-length CDT locus – are characteristic of the majority of variant toxinotypes, the emerging epidemic type BI/NAP1/027 (toxinotype III) being one of the recent well-characterized examples. Interestingly, toxinotype VIII, one of the most widespread variant C. difficile types, characterized by changes in tcdA and tcdB and with an A–B+ phenotype, possessed neither a tcdC deletion nor any form of the CDT locus.
|
ACKNOWLEDGEMENTS |
|---|
|
REFERENCES |
|---|
|
TOPINTRODUCTIONMETHODSRESULTS AND DISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  G. P. Carter, D. Lyras, D. L. Allen, K. E. Mackin, P. M. Howarth, J. R. O'Connor, and J. I. Rood Binary Toxin Production in Clostridium difficile Is Regulated by CdtR, a LytTR Family Response Regulator J. Bacteriol., October 15, 2007; 189(20): 7290 - 7301. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| INT J SYST EVOL MICROBIOL | J MED MICROBIOL | MICROBIOLOGY | J GEN VIROL | ALL SGM JOURNALS |
). Numbers indicate positions in the nucleotide sequence. The positions of two PCR fragments are indicated with solid lines. The CDTc fragment amplifies the entire binary toxin locus, whilst the BTb fragment is generally used for C. difficile-specific detection of binary toxin genes. (b) Amplification of the CDTc PCR fragment in representative toxinotypes, revealing the presence of the complete or truncated CDT locus. M, 500 bp DNA ladder (Gibco Invitrogen); XIa, positive control with functional genes for the binary toxin; 630, positive control with a truncated CDT locus.
